key: cord-1015254-d58yglas
authors: Lee, Chang Heon; Seok, Hyunho; Jang, Woohyuk; Park, Geunsang; Kim, Ji Tae; Kim, Hyeong-U; Rho, Jihun; Kim, Taesung; Chung, Taek Dong
title: Bioaerosol monitoring by integrating DC impedance microfluidic cytometer with wet-cyclone air sampler
date: 2021-07-13
journal: Biosens Bioelectron
DOI: 10.1016/j.bios.2021.113499
sha: 884b5f563e99a110e5aae3473ebf4c3147ec47c6
doc_id: 1015254
cord_uid: d58yglas

The recent outbreak of COVID-19 has highlighted the seriousness of airborne diseases and the need for a proper pathogen detection system. Compared to the ample amount of research on biological detection, work on integrated devices for air monitoring is rare. In this work, we integrated a wet-cyclone air sampler and a DC impedance microfluidic cytometer to build a cyclone-cytometer integrated air monitor (CCAM). The wet-cyclone air sampler sucks the air and concentrates the bioaerosols into 10 mL of aqueous solvent. After 5 min of air sampling, the bioaerosol-containing solution was conveyed to the microfluidic cytometer for detection. The device was tested with aerosolized microbeads, dust, and Escherichia coli (E. coli). CCAM is shown to differentiate particles from 0.96 to 2.95 μm with high accuracy. The wet cyclone air-sampler showed a 28.04% sampling efficiency, and the DC impedance cytometer showed 87.68% detection efficiency, giving a total of 24.59% overall CCAM efficiency. After validation of the device performance, CCAM was used to detect bacterial aerosols and their viability without any separate pretreatment step. Differentiation of dust, live E. coli, and dead E. coli was successfully performed by the addition of BacLight bacterial viability reagent in the sampling solvent. The usage could be further extended to detection of specific species with proper antibody fluorescent label. A promising strategy for aerosol detection is proposed through the constructive integration of a DC impedance microfluidic cytometer and a wet-cyclone air sampler.

biological detection, work on integrated devices for air monitoring is rare. In this work, we integrated 17 a wet-cyclone air sampler and a DC impedance microfluidic cytometer to build a cyclone-cytometer 18 integrated air monitor (CCAM). The wet-cyclone air sampler sucks the air and concentrates the 19 bioaerosols into 10 mL of aqueous solvent. After 5 min of air sampling, the bioaerosol-containing 20 solution was conveyed to the microfluidic cytometer for detection. The device was tested with 21 aerosolized microbeads, dust, and Escherichia coli (E. coli). CCAM is shown to differentiate particles 22 from 0.96 to 2.95 µm with high accuracy. The wet cyclone air-sampler showed a 28.04% sampling 23 efficiency, and the DC impedance cytometer showed 87.68% detection efficiency, giving a total of 24 24.59% overall CCAM efficiency. After validation of the device performance, CCAM was used to 25 detect bacterial aerosols and their viability without any separate pretreatment step. Differentiation of 26 dust, live E. coli, and dead E. coli was successfully performed by the addition of BacLight bacterial 27 viability reagent in the sampling solvent. The usage could be further extended to detection of specific 28 species with proper antibody fluorescent label. A promising strategy for aerosol detection is proposed 29 through the constructive integration of a DC impedance microfluidic cytometer and a wet-cyclone air

The recent COVID-19 pandemic has struck globally and the world is still recovering from its impact.

Pandemic has harassed mankind since the advent of civilization. Millions of people die at every 35 incidence of pandemic and billions of dollars are spent on response and research. Finding a cure for a 36 pandemic not only incurs massive costs but also requires years of research and clinical trials. Until a 37 cure is found, minimizing the spread of pathogens with biosurveillance is the best approach against 38 pandemics. Biosurveillance refers to the systematic monitoring of all data related to biological threats.

This includes continually sampling air or environmental sources and testing them for biological agents 40 (Kman and Bachmann, 2012; Minogue et al., 2019) . The early detection of pathogens in air samples 41 could be used to quarantine people or places to prevent further spread of the diseases. Therefore, an 42 effective detection method for bioaerosols, through which most airborne infections occur, is required 43 J o u r n a l P r e -p r o o f (Kalogerakis et al., 2005; Nazaroff, 2016) . Despite the ample amount of research on air sampling and 1 biological detection methods, only a few have tried to integrate the methods to create bioaerosol 2 detection platforms. Bioaerosol detection has not been standardized or established let alone applicable 3 in the field (Caruana, 2011; Huffman et al., 2020) .

Bioaerosols are airborne collections of biological materials. They are first collected with an air sampler 5 in a liquid state, and the collected liquid sample is then analyzed by microscopy, culture techniques, 6 polymerase chain reaction (PCR), or enzyme-linked immunosorbent assays (ELISA) (Huffman et al., 7 2020; Xu et al., 2011) . However, these methods either require trained operators or take a prolonged 8 period to obtain results, which makes it unsuitable for real-time monitoring (Choi et al., 2014; Ghosh 9 et al., 2015) . For example, PCR, which is the basis for patient diagnosis of COVID-19, requires skilled 10 personnel to work for hours to days to obtain the result (Morales-Narváez and Dincer, 2020). Complex 11 experimental procedures also make it difficult to integrate these methods with air samplers.

Fluorescence-activated cell sorting (FACS) is another widely used method for counting and detecting 13 microorganisms (Chen and Li, 2005; Lange et al., 1997; Prigione et al., 2004) . FACS is intrinsically 14 adequate for real-time detection because it does not require an incubation period and retains sensitivity 15 due to its particle-by-particle analysis principle. However, the high cost and bulkiness of the FACS 16 device is problematic for its use as a biosurveillance method (Joo et al., 2010) . A breakthrough is being 17 made with the introduction of microfluidics-based devices that are small, inexpensive, and easy to 18 integrate with other operational components such as valves, pumps, mixers, and detectors (Ateya et al., 19 2008; Kim et al., 2009; McClain et al., 2001; Shrirao et al., 2018; Whitesides, 2006) . Real-time 20 bioaerosol detection was demonstrated through a micro-optofluidic platform, which was essentially a 21 miniaturization of FACS device (Choi et al., 2015) .

This platform can be further improved by incorporating a DC impedance unit for particle sizing.

Although an optical scatter signal is frequently used for particle sizing, its correlation with particle size 24 is intrinsically non-monotonic (Liu and Daum, 2000; Stier and Quinten, 1998) . This makes it necessary 25 to perform calibration using standard particles, which is still not accurate because of the differences in 26 refractive indices of standard particles and target particles (Rosenberg et al., 2012) . However, the peak 27 magnitude of DC impedance is proportional to the volume of particles regardless of refractive indices 28 (Hurley, 1970; Qin et al., 2011) . The DC impedance detection unit offers more reliable particle size 29 information than the optical scattering unit (Miller and Limes, 1988) . It is also much smaller, cheaper, 30 and requires a less complicated setup. The DC impedance detection unit has been successfully 31 introduced for measuring particle size but has never targeted airborne particles as of yet (Choi et al., 32 2013; Chun et al., 2005; Fu et al., 2017; Guo et al., 2015a; Kim et al., 2009; Ren et al., 2016; Shrirao et 33 al., 2018) .

In this work, we integrated the wet-cyclone air sampler and DC impedance microfluidic cytometer to 35 develop an integrated system for detecting airborne bacteria without any pretreatment step. The cyclone 36 air sampler is frequently used for its high collection efficiency, high flow rate, portability, and 37 compatibility with quantification methods (Kim et al., 2018; Sung et al., 2017 

The resulting solution was redistributed into 50 mL centrifuge tubes and centrifuged at 10,000 rpm for 10 min at 4 °C. The supernatant was removed, and the residue was then resuspended in 40 mL of 1 Dulbecco's phosphate-buffered saline (DPBS; Welgene, Korea) and incubated at 25 °C for 1 h. The 2 incubated sample was centrifuged at 10,000 rpm for 10 min at 4 °C. The supernatant was removed, and 3 the resulting pellet was resuspended in 20 mL of DPBS.

The generation and sampling of bioaerosol were carried out in an air chamber with a volume of 125 L 5 (0.5 × 0.5 × 0.5 m 3 ) as shown in Fig. 1 . To prevent outward spreading during the experiments, a high-6 efficiency particulate air (HEPA) filter was constructed beside the wall of the air chamber. The 2.07 7 μm bead suspension (Bangs laboratory, 20 mL, 5 × 10 7 particles mL −1 ) or bacterial suspension (20 mL, 8 10 7 cells mL −1 ) was aerosolized using a 6-jet collision nebulizer (BGI, USA). Aerosolization was 9 performed using a 4 L min −1 flow of filtered air, which was controlled using a mass flow controller 10 (VICD220; MFC Korea, Korea). The aerosolization time was adjusted by considering the required 11 concentration of aerosol particles. The generated aerosol was dried by a diffusion dryer. Inside the 12 chamber, air fans were used for homogeneity. The air inside the chamber was maintained at 25 °C and 13 a relative humidity of 50% using a thermo-hygrostat. The fabrication process of the microfluidic cytometer was similar to the photolithographic techniques 6 in our previous reports (Choi et al., 2014 (Choi et al., , 2013 . In brief, glass slides (Marienfeld Laboratory Glassware) 7 were cleaned in piranha solution (H2SO4:H2O2 = 3:1) for 30 min, then rinsed with deionized (DI) water, Kenilworth, NJ, USA) was spin-coated onto the surface of the glass slide using a spin coater at 6000 13 rpm for 30 s. The PR-coated slides were placed on a hot plate at 110 °C for 1.5 min and cooled at room 14 temperature for 15 min. The slides were aligned under a film photomask and exposed to UV light (365 15 nm, 18 mJ cm −2 ) for 25 s with a UV aligner (Midas System Co., Ltd., Korea). The UV-exposed slides 16 were immersed in an AZ400K developer (Merck, Kenilworth, NJ, USA) to develop UV-exposed PR.

The glass slides were washed with DI water and dried in clean air. Next, the slides were hard-baked at The simulation of the air sampler at different flow rates is shown in Fig. S1 . A flow rate of 300 L min -40 1 resulted in a cut-off diameter of 0.3 μm, which is small enough to capture particles covering most of 41 the range of bacteria (Levin and Angert, 2015) . Accordingly, the wet cyclone air sampler was operated 42 at a flow rate of 300 L min -1 in all subsequent experiments.

The concentration of particles in the range of 1-3 μm inside the air chamber was measured using an 44 aerosol spectrometer, as shown in Fig. 2 . The aerosolization period is colored in yellow and the air 45 sampling period is colored in orange. Particles were observed for 15 min, and for the first 10 min, 46 different solutions were aerosolized. When DI water was aerosolized, there was no change in the J o u r n a l P r e -p r o o f concentration of 1-3 μm particles for the entire 10 min, as shown in Fig. 2(a) . Fig. 2(b) shows 2.07 μm 1 beads in DI water aerosolized for 10 min. A noticeable increase in the particle count indicates that the 2 2.07 μm beads are aerosolized as expected. Fig. 2(c) shows the case of 2.07 μm beads aerosolized for 3 10 min and subsequently air-sampled for 5 min. The graph shows that the beads were effectively 4 eliminated from the air. A similar approach was adopted for E. coli, but the spectrometer signal from 5 DPBS was so high that the signal from E. coli was undistinguishable from the DPBS signal, as shown 6 in Fig. 2(d) . 

A two-channel microfluidic cytometer was designed and fabricated to detect bioaerosols collected from 4 an air sampler, as shown in Fig. 3 . The cytometer was based on our previous design and was improved 5 to serve our purpose (Choi et al., 2014; Joo et al., 2010) . A syringe pump drives the sample solution 6 from the wet cyclone air sampler into the microfluidic cytometer. The sample solution injected through 7 the inlet reservoir passed through a split channel where impedance and fluorescence sensing were 8 performed. Each microfluidic sensing channel was 30 µm in length and 10 µm in width and depth. At 9 the end of each channel, the Ag/AgCl electrode was located for ionic current measurement, as shown 10 in Fig. 3(b) . A constant DC voltage of 0.8 V was applied between the two Ag/AgCl electrodes, and the 11 chloride ion flow produced an ionic current. As particles pass through the sensing region, they block 12 the narrow channel interfering with chloride flow, resulting in a decrease in ionic current, causing an 13 impedance peak to appear. In principle, the impedance peak amplitude should be linear to the volume 14 of the particle (Choi et al., 2013; Chun et al., 2005; Guo et al., 2015b; Kim et al., 2009; Rho et al., 2018) .

At the same time, blue light irradiates the fluorescence detection region shown in Fig. 3(a) 

The correlation between the impedance peak amplitude and microbead volume was assessed with 31 aerosolized fluorescent microbeads of different diameters (0.96, 2.07, and 2.95 μm). The actual 32 impedance data is presented in Fig. S2 . The peak amplitudes and particle volumes obtained using the 33 newly designed cytometer are plotted in Fig. 4(a) . The calibration curve with a linearity of R 2 = 0.9967 34 proves that the principle fits well with this cytometer. A calibration curve was used to draw the 35 histogram in Fig. 4(b) . The diameters of the microbeads calculated from the peak intensities were clearly 36 J o u r n a l P r e -p r o o f distinguishable from each other. E. coli BL21 was also measured, and the calculated diameters were 1 plotted along with the beads in Fig. 4(b) . The diameters and statistical calculations of the beads and E.

2 coli are presented in Table S1 . E. coli has a cylindrical structure with a hemispherical cap with a 3 diameter of approximately 1 μm and a length of approximately 2 μm. The mean diameter of 1.641 μm 4 calculated in Table S1 seems to be in good agreement with the actual value. A rather large standard 5 deviation compared to those of beads comes from the actual inconsistencies in E. coli cell sizes (Yao et 6 al., 2012) . The cytometer was also used to deduce bead concentration successfully as presented in The efficiency of the bioaerosol detection system was evaluated with aerosolized 2.07 μm beads, as 16 shown in Fig. 5(a) . Beads were aerosolized for 1, 3, 5, 7, and 10 min to make different concentrations 17 in the chamber, and the concentrations were measured with the spectrometer. Then, the air sampler 18 collected the aerosolized beads into 10 mL DPBS solution for 5 min. Part of the solution was put through 19 a DC impedance microfluidic cytometer and another part was measured with a BD FACS Canto™ II 20 as a reference. Fig. 5 (a) compares the total bead counts calculated from the three different measurements.

Comparison of particle counts from the spectrometer and FACS elicited collection efficiency. The 22 collection efficiency varied depending on the concentration and was 28.04% on average. This is because 23 of two factors, the inaccuracy of aerosol spectrometer and sedimentation/adsorption of aerosol during 24 collection. First, aerosol spectrometer measures all particles including water droplets. Therefore, 25 overestimation is expected. Secondly, sedimentation occurs as can be seen in Fig. 2 J o u r n a l P r e -p r o o f measurements, which was 24.59% on average. The device could be said to perform efficiently over the 1 tested particle concentration range.

The device efficiency was also observed in E. coli. E. coli was diluted in DPBS solution when 3 aerosolized to preserve the structure of E. coli. Because of the dried salts produced from the DPBS 4 solution, the concentration of particles in the air could not be measured accurately, so the spectrometer 5 data were missing from the E. coli measurement. Fig. 5(b) compares the results from FACS and the 6 microfluidic cytometer. The concentrations of E. coli solution measured by the microfluidic cytometer 7 were, on average, 81.61% compared to the concentration measured by FACS. The reason for the low 8 concentration of bacteria compared to microbeads might be (1) E. coli tends to adhere to the wall of the 9 microfluidic channel and (2) the diameter of E. coli was smaller than that of beads and might be less 10 efficient at the collection step. For these reasons, the data measured here support the usefulness of 11 combining the cyclone sampler with a microfluidic cytometer for reasonably fast detection of the target 12 microbial cells in bioaerosols. (Iswardy et al., 2017; Lippe, 2018 We constituted the first air monitoring system to utilize DC impedance for particle size measurement. 

The good, the bad, and the tiny: a review of microflow cytometry

Detection and analysis of airborne particles of biological origin: present and future

Sampling performance for bioaerosols by flow cytometry with fluorochrome

A flow cytometry-based submicron-sized bacterial detection system using a movable virtual wall

A label-free DC impedance-based microcytometer for circulating rare cancer cell counting

Integrated micro-optofluidic platform for real-time detection of airborne microorganisms

Cytometry and velocimetry on a microfluidic chip using polyelectrolytic salt bridges

Lab-on-PCB-based micro-cytometer for circulating tumor cells detection and enumeration

Review of bioaerosols in indoor environment with special reference to sampling, analysis and control mechanisms

Volumetric measurement of human red blood cells by MOSFET-based microfluidic gate

Numerical and experimental characterization of solid-state micropore-based cytometer for detection and enumeration of biological cells

Real-time sensing of bioaerosols: Review and current perspectives

Sizing Particles with a Coulter Counter

A bead-based immunofluorescence-assay on a microfluidic dielectrophoresis platform for rapid dengue virus detection

A portable microfluidic flow cytometer based on simultaneous detection of impedance and fluorescence

Indoor air quality-bioaerosol measurements in domestic and office premises

Electrochemical Detection of Airborne Influenza Virus using Air Sampling System

Red blood cell quantification microfluidic chip using polyelectrolytic gel electrodes

Biosurveillance: A Review and Update

Application of flow cytometry and fluorescent in situ hybridization for assessment of exposures to airborne bacteria

Small but mighty: Cell size and bacteria

Flow Virometry: a Powerful Tool To Functionally Characterize Viruses

THE EFFECT OF REFRACTIVE INDEX ON SIZE DISTRIBUTIONS AND LIGHT SCATTERING COEFFICIENTS DERIVED FROM OPTICAL PARTICLE COUNTERS

Flow cytometry of Escherichia coli on microfluidic devices

Recent advances in particle size measurements: A critical review. C R C

Next-generation sequencing for biodefense: Biothreat detection, forensics, and the clinic

The impact of biosensing in a pandemic outbreak: COVID-19

Indoor bioaerosol dynamics

Development and Use of Flow Cytometry for Detection of Airborne Fungi

Effects of particle's off-axis position, shape, orientation and entry position on resistance changes of micro Coulter counting devices

Lab on dielectric film deposited PCB device for characterization of electrical property of biological cells

Multiplex immunoassays using virus-tethered gold microspheres by DC impedance-based flow cytometry

Particle sizing calibration with refractive index correction for light scattering optical particle counters and impacts upon PCASP and CDP data collected during the Fennec campaign

Microfluidic flow cytometry: The role of microfabrication methodologies, performance and functional specification

Simple refractive index correction for the optical particle counter PCS 2000 by Palas

Highly efficient in-line wet cyclone air sampler for airborne virus detection

The origins and the future of microfluidics

Bioaerosol science, technology, and engineering: Past, present, and future

Regulation of cell size in response to nutrient availability by fatty acid biosynthesis in Escherichia coli