Recent progress in online detection methods of bioaerosols

Highlights

• •Online bioaerosols detection is still challenging due to their complexity and variability.
• •ATP bioluminescence is a cheap, rapid and near real-time monitoring method.
• •Newer UV-LIF instruments can distinguish between bioaerosol types.
• •SERS and CARS are also rapid and specific technologies to discriminate bacterial spores.
• •BAMS and MALDI-TOF MS has higher accuracy and sensitivity for bioaerosol, respectively.

Abstract

The aerosol transmission of coronavirus disease in 2019, along with the spread of other respiratory diseases, caused significant loss of life and property; it impressed upon us the importance of real-time bioaerosol detection. The complexity, diversity, and large spatiotemporal variability of bioaerosols and their external/internal mixing with abiotic components pose challenges for effective online bioaerosol monitoring. Traditional methods focus on directly capturing bioaerosols before subsequent time-consuming laboratory analysis such as culture-based methods, preventing the high-resolution time-based characteristics necessary for an online approach. Through a comprehensive literature assessment, this review highlights and discusses the most commonly used real-time bioaerosol monitoring techniques and the associated commercially available monitors. Methods applied in online bioaerosol monitoring, including adenosine triphosphate bioluminescence, laser/light-induced fluorescence spectroscopy, Raman spectroscopy, and bioaerosol mass spectrometry are summarized. The working principles, characteristics, sensitivities, and efficiencies of these real-time detection methods are compared to understand their responses to known particle types and to contrast their differences. Approaches developed to analyze the substantial data sets obtained by these instruments and to overcome the limitations of current real-time bioaerosol monitoring technologies are also introduced. Finally, an outlook is proposed for future instrumentation indicating a need for highly revolutionized bioaerosol detection technologies.

Graphical abstract

1. Introduction

Bioaerosols include live and dead bacteria, fungi, and viruses, their excretions (e.g., endotoxin, glucans, and mycotoxins), fungal spores, and plant pollen; they are recognized as one of the main routes for microbial dispersal and transmission [1,2]. They can be formed from various means such as talking, coughing, sniffling, walking, taking a bath, flushing the toilet, and sweeping the floor, as well as through industrial activities [3]. Various airway diseases, including rhinitis, airway inflammation, asthma, organic dust toxic syndrome, seasonal influenza, severe acute respiratory syndrome, and coronavirus disease (COVID-19), are related to bioaerosol exposure [4,5]. For example, tuberculosis, which is the leading global infectious cause of death in adults, is transmitted by airborne Mycobacterium tuberculosis [6]. Inhalation of high concentrations of Aspergillus fumigatus is associated with invasive aspergillosis or allergic bronchopulmonary aspergillosis [7]. Allergic rhinitis (hay fever) and asthma have been related to elevated airborne Artemisia and Humulus pollen concentrations [8]. Talking, laughing, singing, coughing, sneezing, or direct contact with the mouth, nose, or eyes can lead to the spread of the SARS-CoV-2 virus from infected persons to others, and has caused more than 6 million deaths globally [9]. The above phenomena have warned us of the importance of real-time bioaerosol monitoring for airborne pathogens. To date, studies of airborne pathogen detection have been undertaken with varying success.

Detection methods for bioaerosols include traditional offline and online measurement techniques [10]. Traditional methods are based on bioaerosol sample collection and subsequent laboratory analysis, such as plate culture colony counting, microscopy, and polymerase chain reaction (PCR) [11,12]. These techniques are limited in their ability to yield the high-resolution time series of data that is necessary for an online approach. Culture-based colony counting is the most widely used method for bioaerosol detection [13], but is only suitable for culturable microorganisms. It is also time-consuming [14] and may exclude viable but non-culturable cells that could become culturable under suitable conditions [15], resulting in an underestimation of the bioaerosols [16]. Bioaerosols captured on a filter and resuspended in phosphate buffer saline can be quantified using DNA-based molecular techniques [17] and microscopy [18,19]. Quantitative PCR (qPCR) is a sensitive, specific method that enables rapid quantification of microorganisms of interest and their key functional genes; however, the results depend on DNA extraction efficiency and primers [20]. Direct microscopic counting methods enable the enumeration of total bioaerosols but require samples with high concentration, and are time-consuming due to cumbersome sample post-processing requirements such as staining and repeated filtration [21]. Because of this, these techniques cannot be used for real-time detection, as they lack adequate sampling volume, rapid operation, and nondestructive microorganism analysis.

In contrast, the adenosine triphosphate (ATP) bioluminescence assay, laser/light-induced fluorescence (LIF) spectroscopy, Raman spectroscopy, and bioaerosol mass spectrometry (BAMS) [11] allow for high resolution real-time characterization, and were developed to achieve online detection of biological aerosols. Moreover, breath-borne volatile organic compounds (VOCs) or protein biomarkers from rats have also been used to immediately warn about toxic air for real-time monitoring of particle toxicity [22,23]. However, each method differs in terms of its working principles, performance, sensitivity, accuracy, and efficiency. Additionally, online detection of bioaerosols is still challenging due to the complexity and variability of pathogenic microorganisms present. This review mainly provides insight into the existing technologies for online monitoring of bioaerosols. The main focus of the paper is to review different real-time bioaerosol monitoring methodologies, the present status of technological advances, and the major developments and applications in bioaerosol research. Furthermore, recent advances and future perspectives of bioaerosol identification are emphasized. Overall, this review is expected to provide improved knowledge for selecting appropriate methods for bioaerosol monitoring, acting as an asset to numerous fields of research.

2. Adenosine triphosphate bioluminescence assay

ATP bioluminescence quantifies microorganisms by measuring the amount of fluorescence produced during ATP's reaction with luciferase and luciferin, as the emitted light intensity is directly proportional to the ATP concentration in the biomass (Fig. 1) [24]. As early as the 1950s, National Aeronautics and Space Administration (NASA) scientists proposed the ATP bioluminescence assay to identify living cells on other planets [25,26]. In the following decades, ATP bioluminescence became rapidly and widely employed to monitor microbial pollution on target surfaces in the food and healthcare industries in combination with a swab-based commercial luminometer that provided results within a short time (< 1 min) [24]. Because it was simple, cost-effective, and rapid, it has been used to determine the surface cleanliness of hospital kitchens and rapidly monitor airborne microorganisms [27,28]. To achieve real-time monitoring with this technique, however, bioaerosol samplers need to be easily compatible with the ATP bioluminescence assay. A filter-based sampler, an electrostatic sampler, and a bioaerosol-to-hydrosol sampler (Fig. 2) have already been employed with the ATP-based assay, yielding satisfactory results [24,29].

Fig. 1.

The chemical reaction theory of adenosine triphosphate (ATP) bioluminescence assay. AMP: adenosine monophosphate.

Fig. 2.

ATP bioluminescence assay based on different bioaerosol samplers. (a) swab (from ref. [31]); (b) automated bioaerosol-monitoring system (from ref. [38]); (c) electrostatic sampler with swab (from ref. [41]); and (d) electrostatic sampler with liquid collected unit (from ref. [42]).

2.1. Impactor and filter-based samplers

Impactors and filters have been coupled with commercial swab-based luminometers for real-time monitoring of bioaerosols. They collect bioaerosols onto solid/semi-solid media, high-efficiency particulate air (HEPA) filters, or filters made of polycarbonate, mixed cellulose esters, Teflon, polytetrafluoroethylene, nylon, polyvinyl chloride, and gelatin to make sampling relatively simple, less expensive, and highly effective [2,30]. As shown in Fig. 2a, under the control of an Arduino-Bluetooth smartphone controller, near real-time (< 3 min) monitoring of both aerosolized Staphylococcus aureus and field bioaerosols was achieved by efficiently extracting ATP from bioaerosols captured in microfiber swabs and amplifying the bioluminescence signal (> 10⁴ relative luminescent units (RLU)/m³ air) [31]. Yoon et al. used an inertial impactor (with a 1 µm cutoff diameter) with an ATP bioluminescence detector to estimate the efficacy of air-controlling devices for bioaerosol removal in situ within 25 min; linearity was obtained between this method and the culture-based method in both laboratory and field tests (R² = 0.9283 and 0.8189, respectively) [24]. Park et al. found that when using a carbon fiber ionizer for lysing airborne bacteria collected on glass microfiber membrane filters, the ATP bioluminescence detection system can achieve real-time monitoring of bioaerosols in under 4 min [32]. Overall, ATP bioluminescence assays in combination with impactors and filters cannot be easily applied for continuous and online bioaerosol monitoring because time is unavoidably required for sample preparation and other manual procedures. Additionally, the loss of viability due to desiccation might result in an underestimation of bioaerosol concentration, particularly for sensitive organisms or agents [30,33].

2.2. Liquid-based samplers

Impingers and cyclones such as the Coriolis sampler and SKC QuickTake 30 sampler can overcome the deficiencies of impaction processes by collecting airborne particles into a liquid medium [34]. In this way, the rebound effect of particles is reduced and the activity of microorganisms is preserved [35]. These have been coupled with an ATP luminometer for rapid monitoring of indoor bioaerosols. For example, Lee et al. used an aerosol condensation system to produce hydrolyzed samples, demonstrating airborne Bacillus subtilis and Escherichia coli JM110 could be detected in real-time within 10 min in a downstream transducer after extracting the intracellular ATP in a microfluidic channel [36]. Heo et al. developed an aerosol-to-hydrosol sampler with a super hydrophilic surface by coating silica nanoparticles to the inner surface of a wet-cyclone to continuously enrich bioaerosols (increased 2.4 × 10⁶ fold); the bioaerosols then flow to a microfluidic channel to measure ATP bioluminescence using an optofluidic bioluminescence detector with a detection limit > 50 colony-forming units (CFU)/m_air³ [37]. As shown in Fig. 2b, Cho et al. constructed an automatically operated on-site bioaerosol-monitoring system for airborne E. coli with detection intervals of 5 min, a detection limit of ∼130 CFU/m_air³, and a high enzyme reaction stability of > 30 days. It consisted of a wet-cyclone sampler for sampling bioaerosols, a thermal-lysis unit for extracting ATP, and an ATP bioluminescence-detection unit with luciferase-luciferin immobilized in a glass-fiber conjugation pad [38]. In general, the liquid-based collection method can maintain high cell viability, but its relatively low sampling efficiency and enrichment performance at a high flow rate need to be addressed in future studies.

2.3. Electrostatic-based sampler

An electrostatic sampler was developed to collect charged bioaerosols as this method has lower impaction stress, pressure drops, and power consumption in comparison to impactors and has a higher collection efficiency than impingers [39,40]. As Fig. 2c shows, to achieve fast bioaerosol monitoring within 4 min, Park et al. first constructed a novel hand-held electrostatic rod-type sampler comprising a charger and a collector; these were placed in conjunction with a commercial luminometer to test concentrations of airborne S. epidermidis and indoor bioaerosols [41]. They further developed a single-stage electrostatic sampler for continuous and real-time monitoring of bioaerosols (Fig. 2d). The bacterial aerosols were charged by atmospheric ions and sampled in a flowing liquid containing both lysis buffer and d-luciferin-luciferase reagent and were then measured in a bioluminescence detector with a total response time of 30 s at a liquid sampling flow rate of 800 µL/min [42].

In summary, impactors, microfiber swabs, glass microfiber membrane filters, impingers, wet-cyclones, and electrostatic samplers have been combined with the ATP bioluminescence assay for evaluating sampling performance as well as for rapid and real-time/near real-time bioaerosol monitoring, yielding satisfactory results [29]. However, since the ATP level in spores is very low and viruses do not have ATP, these ATP-based detection technologies perform poorly in detecting spores and are not suitable for viral analysis [43]. Moreover, they cannot identify specific bacteria/fungi species because ATP is present in all living bacteria and fungi [44]. The sensitivity of this method is also affected by environmental factors, as bioaerosols can lose their viability under physical stress or exposure to harsh environments [38]. Finally, although advanced detection systems such as automated bioaerosol-monitoring systems can be used for continuous and real-time surveillance of bioaerosols, the technology is still developing and is therefore not yet available for real-world application as some specifics such as cost are still unknown.

3. Ultraviolet laser/light-induced fluorescence

The well-established offline measurement of fluorescent materials by microscopy provided the basics for the application of fluorescence in real-time bioaerosol measurement. Significant attention has been paid in the last decades to developing online ultraviolet laser/light-induced fluorescence (UV-LIF) instruments because they allow non-invasive and near-instantaneous detection of particles [45]. As shown in Fig. 3, UV-LIF instruments are based on the principle that primary biological aerosol particles (PBAPs) contain biofluorophores, such as NAD(P)H (NADH and NADPH), riboflavin, and amino acids (histidine, tryptophan, and phenylalanine) that auto-fluoresce under UV light/laser exposure; the excitation and emission wavelengths can be used to detect biological fluorophores in bioaerosols [46]. This method was originally developed by a military group to rapidly detect biological warfare agents (BWAs), such as pathogens, viruses, and toxins [47]. Instruments that can reliably monitor BWAs with high accuracy and fast response, even at low concentrations, include the fluorescent aerodynamic particle sizer (FLAPS), ultraviolet aerodynamic particle sizer (UV-APS), and biological aerosol warning system (BAWS). FLAPS is the most prominent detector and has an additional helium-cadmium laser that can measure the particle size and fluorescence in the air. UV-APSs are a variant of the FLAPS that nonspecifically detect biological agents by measuring particle size and fluorescence intensity. BAWSs are micro-laser devices that can detect and discriminate biological particles from other natural airborne particles in real-time [48]. Two decades of research on detecting BWAs have promoted the recent development and commercialization of UV-LIF instruments. These LIF instruments include the UV-APS, wideband integrated bioaerosol sensors (WIBS), BioScout, and the newly-developed LIF-based instruments such as the spectral intensity bioaerosol sensor (SIBS) and Rapid-E that have been developed in the past few decades [49].

Fig. 3.

The chemical theory of ultraviolet laser/light-induced fluorescence. The excitation (Ex) and emission (Em) wavelengths of these biofluorophores are cited from ref (from ref. [52]).

3.1. Ultraviolet aerodynamic particle sizer

One of the most commonly used LIF instruments for evaluating individual particle size, concentration, and total fluorescence intensity is the UV-APS (TSI Inc., USA). This device records the fluorescence of individual bioaerosols in the wavelength range of 420–510 nm after excitation by a third ultraviolet laser (Nd:YAG) at 355 nm from a light detection and ranging (LiDAR) detector [50]. It is the first commercially available single-particle LIF-based, real-time sensor with both high temporal and size resolution [49]. The first wave of UV-APS studies in the laboratory sought to evaluate its performance. Agranovski et al. evaluated the selectivity, counting efficiency, sensitivity, and limit of detection of the UV-APS using atomized bacterial and non-bacterial particles, and found that the counting efficiency depended on the concentration. The instrument had an upper detection limit of 6 × 10⁷ particles/m³; it was capable of detecting NADH and riboflavin, but not NADPH, likely because the amount of NADPH in most bacteria is approximately five times lower than the amount of NADH [51,52]. Since NADH/NADPH and riboflavin are found in metabolically active cells, the UV-APS could also be used to specifically distinguish viable bioaerosols [53]. Subsequent research assessed the performance of bioaerosol samplers (the six-stage Andersen microbial impactor, SKC biological sampler, AGI-30 impingers, and others), real-time bioaerosol detectors, and cleaning appliances [54,55]. Specifically, the concentration of fluorescent particles detected by the UV-APS was more than one order of magnitude higher than that detected by AGI-30 impingers [52]. By investigating the particles and bacteria emitted from 21 vacuums using UV-APS, Knibbs et al. found that emitted particles varied significantly from 4.0 × 10⁴ to 1.2 × 10⁹ particles/min [54]. Examining the influences of fungal age on their fluorescence and physical properties, Kanaani et al. found that particle size increased with age, while the fluorescent percentage of spores decreased [56]. Further examining the effect of simulated ozone on bioaerosols, Roshchina and Karnaukhov found that after a 100-h exposure, the maximum fluorescence at 530–550 nm disappeared, shifting to 475–480 nm for Philadelphus grandiflorus and Epiphyllum hybridum; this suggested ozone may have caused damage to the pollen [57].

As the first commercial online bioaerosol monitoring instrument, although the UV-APS has low sensitivity and accuracy toward airborne microorganisms such as fungal spores, it still has a place in the online detection of bioaerosols as it can adapt to a variety of weather conditions. However, as the environmental conditions will affect UV fluorescence imaging in the field, the LIF instrument should be operated only under suitable temperatures and humidity [58].

3.2. Wideband integrated bioaerosol sensor

Another widely-used instrument for monitoring bioaerosols in real-time is the WIBS, which was invented by Professor Paul Kaye at the University of Hertfordshire, UK, and commercialized by Droplet Measurement Technologies, USA. There are two WIBS prototypes (WIBS-3 and WIBS-4) and two commercially available models (WIBS-4A and WIBS-5/NEO). The WIBS-4 is improved from the WIBS-3 in terms of optical chamber design, the optical filters employed for detecting wavebands, UV delivery, and particle sizing [59]. The WIBS-4 prototype measures particle sizing using a dual gain detection approach; the high gain is used for monitoring smaller size particles (0.5–12 µm) and the low gain detects larger size particles (3–31 µm). The commercialized WIBS-4A only has a single gain setting, which measures a single particle size (0.5–20 µm), asymmetry (shape), and fluorescence intensity in three channels by evaluating forward and lateral light scattering and the unresolved spectral fluorescence intensity of each particle at millisecond resolution. These fluorescence channels include FL1 (excitation/emission (Ex/Em) = 280/310–400 nm), FL2 (Ex/Em = 280/420–650 nm), and FL3 (Ex/Em = 370/420–650 nm) (Fig. 4). Scattering of the laser at 635 nm is used for particles sizing using the Mie theory, while aerosol morphology is determined through asymmetry factor (AF) measurement [60].

Fig. 4.

Schematic diagram of the WIBS detection showing the orientation of the Xenon flash lamps, PMT detectors, red triggering laser, and spherical mirrors. FL1 detector recorded the fluorescence of FL1 channel (Ex/Em = 280/310–400 nm); while FL2 detector recorded the fluorescence of FL2 channel (Ex/Em = 280/420–650 nm) and FL3 channel (Ex/Em = 370/420–650 nm) (from ref. [67]).

Similar to the UV-APS, most studies have focused on evaluating the performance of WIBSs in the laboratory. Healy et al. found that the WIBS-4’s counting efficiencies were 50% and 100% for polystyrene latex (PSL) spheres with diameters of 489 and 690 nm, respectively, suggesting that when using it to analyze particles ≤ 690 nm, the counting efficiency must be adjusted [61]. They also found that A. fumigatus, Alternaria alternate, and Penicillium notatum could be distinguished from four other fungal spores, while grass pollen was distinguishable from seven other pollens based on size using FL1 (Ex/Em = 280/310–400 nm), FL2 (Ex/Em = 280/420–650 nm), FL3, (Ex/Em = 370/420–650 nm) and AF [59]. Toprak et al. found that the combination of two fluorescent channels, FL1 and FL3, made biological and abiotic aerosols (polycyclic aromatic hydrocarbons, ammonium sulfate, and mineral dust) distinguishable [62]. Using WIBS-4A, Robinson et al. detected size-selected particles containing fluorophores of known mass using mixed particles of ammonium sulfate-tryptophan and quinine particles to calibrate the FL1 and FL2 channels [63]. Hernandezn et al. compiled an aerobiological reference catalog using the WIBS-4A that visualizes the size, shape, and fluorescent channel emission intensities of 14 bacterial, 29 fungal, and 12 pollen aerosols; they were able to discriminate between airborne microorganisms and pollen [64]. Using this categorization system, Zhou et al. found that the average emission rate of 1–10 µm fluorescent particles from human walking ranged from 6.8 to 7.5 million particles per person/h [65]. Savage et al. investigated the fluorescence threshold of 69 type aerosols, including bioaerosols and non-biological interferents (smoke, soot, and humic-like substances), and found that these particles could be broadly separated directly using five parameters: three fluorescence channels (FL1, FL2, and FL3), size, and AF [66] along with a Perring-style particle classification scheme [67]. They also found that increasing the fluorescence threshold to FT + 9σ can significantly reduce interference from mineral dust and other abiotic factors [47]. Savage and Huffman developed a hierarchical agglomerative clustering method that improves discrimination between biological and abiotic particles, allowing particle separation by more than an order-of-magnitude level and with < 5 % misclassification [66].

The above studies focused on evaluating the counting efficiency, increasing differentiation sensitivity between bioaerosols and interfering particles, and calibrating fluorescence detection. Compared to UV-APSs (Table 1), WIBSs can utilize multiple flash lamps as excitation sources; besides using 370–405 nm excitation to excite NAD(P)H and flavin in a similar manner to UV-APS, WIBSs can also use 280 nm to excite amino acids, allowing fungal spore detection. WIBSs are more suited for bacterial detection than UV-APS as bacteria fluoresce almost exclusively in the FL1 channel with little fluorescence in F3 channels; this is because the excitation intensity from xenon flash lamps (Xe2) is weak and the rapid sequence of xenon flash lamp excitation leads to quenching of the FL3 channel fluorescence [68]. WIBSs are also the only instrument that has been successfully used to detect laboratory and outdoor environment pollen, and they can distinguish between different fungal spores as well as spores and pollen grains; however, species-level PBAP identification remains to be established.

Table 1.

The instrumental parameters of LIF-based instruments.

Device	Excitation (nm) and/or scatter source	Fluorescence detection range (nm)	Channel	Sizing method	Size range(µm)	Time resolution	Sample flow (L/min)	Maximum Concentration
WIBS-4/4A	280/370 nm Xe flash lamps	310–400 and 420–650	3	Optical diameter	WIBS-4:0.5–12; HG; 3–30 LG; WIBS-4A:0.5–20	ms	WIBS-4: 0.24; WIBS-4A:0.3	2 × 104 (particle/L)
UV-APS	355-nm UV laser	430–580	1	Aerodynamic diameter	0.5–20	1 s–18 h (5 min generally)	1	/
BioScout	405 nm laser diode	> 442	16	Optical diameter	0.5–10	1 s	2	/

3.3. BioScout

The BioScout is a new real-time bioaerosol monitoring system developed and commercialized in Finland [69]. It uses a 405 nm laser diode to induce autofluorescence (mainly from tryptophan, riboflavin, and ovalbumin) from a single bioaerosol [69]. It records autofluorescence across 16 emission wavelength channels (> 422 nm) and uses scattered light to examine the optical particle size; the particles measured are 0.5–10 µm at a 1 s time resolution [49].

Research on the BioScout has mainly focused on comparative studies with UV-APS. For example, Saari et al. found that the BioScout measured higher fluorescent particle fraction (FPF) values (0.34–0.77) for aerosolized bacteria than the UV-APS system (0.13–0.17). For coarse and fine fluorescent biological aerosol particles, the detection efficiencies were 2.6 and 9.7 times higher than for the UV-APS, which was possibly due to the high intensity of the excitation laser and signal-to-noise ratio of the BioScout [70]. For fungal species, comparing culture media, cultivation time, and air exposure velocity, Saari et al. found that the FPF values detected by the BioScout were higher than those of the UV-APS for four-month-old Aspergillus versicolor, but were lower for four-month-old Penicillium brevicompactum. These variations may be due to the compounds with different fluorescence that were produced during the late stage of the spores [71]. For NADH, higher FPF values were detected by UV-APS, which was attributed to its high excitation peak at 355 nm [69].

There are several differences between the UV-APS (Table 1) and the BioScout: The BioScout has a higher counting efficiency for small particles, but a lower particle size resolution. BioScout also cannot detect fungal spores and pollen with sizes > 10 µm; both systems can be used to detect bacteria in the laboratory, but the BioScout has a greater fluorescent fraction. The UV-APS was better for detecting live or metabolically-active biological particles by inducing NAD(P)H fluorescence using a 355 nm laser, while the BioScout was more sensitive for detecting bacterial cells.

3.4. Newer laser/light-induced fluorescence instruments

New real-time bioaerosol monitoring instruments continue to be developed, with much progress being made in increasing the sensitivity and resolution of the fluorescence spectra. Such instruments include SIBS, particle analyzer, multi-parameter bioaerosol sensor (MBS), and Rapid-E. Similar to the above-discussed LIF instruments, these devices can utilize the full fluorescence emission range to help identify fluorescent aerosol particles (FAPs)/PBAP by determining the bioaerosol's spectral fingerprints.

The first type of instrument is WIBS-New Electronic Option (WIBS-NEO/5), developed and manufactured by Droplet Measurement Technologies. It is an upgrade of the WIBS-4A utilizing new operating software [72]. It has been reported as the world's only instrument for single particle measurement of bioaerosols in real-time using two excitation and emission bands [73]. It can specifically distinguish grass pollen from tree pollen by detecting chlorophyll-a fluorescence, which is a possible grass pollen biomarker [74]. The main differences between the WIBS-NEO and previous WIBS instruments have been previously detailed [75] and are summarized in Table 2. Briefly, it provides two additional fluorescence data channels, FL4 [Ex/Em = 280/(600–750) nm] and FL5 [Ex/Em = 370/(600–750) nm], and has a lower sheath flow rate (2.1 L/min) and a broader size range (0.5–30 µm); it also has a larger fluorescence intensity range (∼1 × 10⁴ to ∼2.1 × 10⁹) and presents higher fluorescent intensities for abiotic particles than some bioaerosols due to the detector configuration.

Table 2.

The instrumental parameters of newer LIF-based instruments.

Devices	Excitation and/or scatter source	Fluorescence detection range (nm)	Channels	Sizing method	Size detection range (µm)	Time resolution	Sample flow (L/min)	Maximum counts
PA-1000	263 nm UV laser	390–600	32	Optical diameter	0.5–100	ms	2.8	/
Rapid-E plus	337/445 nm UV laser	380–580	32	Optical diameter	0.3–100	1 ms	5	1 × 106 particle/L
MBS	280 nm Xenon flash lamp	300–655	8	Optical diameter	1–20	5 ms	0.3	/
SIBS	280 and 370 nm, Xenon flashtube	302–721	16	Optical diameter	0.3–30	ms	0.3	2 × 104 particle/L
WIBS-5/NEO	280 and 370 nm Xenon flashtube	310–400 and 420–650	3	Optical diameter	0.5–30	/	0.3	Fluorescent particles: 466 particle/cm3; Total particles: 9500 particle/cm3

The second types of instrument are MBS and SIBS, manufactured by the University of Hertfordshire [46]. MBS is an upgrade of the WIBS technology that can provide the size (1–20 µm), shape, and fluorescence intensity of air particles. The total measurement time of particles using the MBS is 30 µs; it can count particles at a rate > 1000/s; however, the long recharging time of the xenon lamp (5 ms) reduces the acquisition rate of data to about 100 particles/s [76]. In contrast to the WIBS, the MBS presents considerable differences [76]. It records the fluorescence in eight emission bands (from approximately 300 to 655 nm at Ex = 280 nm), which can better distinguish biological particles from interferents; the MBS also uses more detectors to record the morphology from the particle's spatial light scattering pattern, which enhances the particle classification probability and accuracy. The MBS presents higher shape values for particles as well, especially for pollen (Table 2).

The SIBS is comprised of an optical chamber, two pulsed xenon UV sources, a 785 nm diode laser, an avalanche photodiode, a quadrant photomultiplier tube, and 16 channel photomultiplier spectrometers. The SIBS samples bioaerosol at 0.3 L/min and derives the equivalent optical diameter and shape (Table 2). In contrast to the WIBS, the SIBS can measure fluorescence emission across 16 wavelength bands from a wider range of emission wavelengths (Em = 300–720 nm), providing greater spectral resolution for bioaerosols; it has a lower size limit (0.3 µm), which expands the analytical scope and can distinguish chlorophyll a and b from bacteriochlorophyll, providing the possibility of detecting algae, plant residues, and cyanobacteria [77].

The third type is particle analyzers, such as the PA-300, PA-1000, Rapid-E, and Rapid-E Plus produced by Plair SA in Geneva, Switzerland. The PA-300 can identify bioaerosols at a flow rate of 2 L/min by obtaining nearly complete spectra for each particle using light scattering and spectral resolution fluorescence [78]. Particles first pass through a 658 nm continuous wavelength laser to characterize their size and surface morphology. In addition, the PA-300 can record the phosphorescence of the particles using the two photo-detectors, suggesting that it can distinguish between biological species. Moreover, the size (1–100 µm) measured using the optical particle sizer is larger than all three instruments mentioned above [79,80]. Its updated version, PA-1000, can detect aerosol particles as low as 0.5 µm and be used to distinguish between particles with similar physical and spectral characteristics [49]. The Rapid-E is an intelligent bioaerosol sensor that can detect particles from 0.3 to 100 µm at a flow rate of 3 L/min [81]. It utilizes a 337 nm laser to excite particles and 32 channels to identify particles. The light is scattered at 24 angles and can be used to determine particle shape in real time, while the rate of fluorescence decay can be used to identify protein-bearing particles [82]. By comparing the Rapid-E with the WIBS-NEO, Lieberherr et al. found that the Rapid-E performed best for larger particles (PSL, 10 µm) with an efficiency of 58%, while the WIBS-NEO achieved a 90% efficiency for smaller particles (PSL, 0.9 µm) [72]. The Rapid-E also has a higher sample volume (3 L/min) and has been mainly used to classify pollen taxa. Recently, the Rapid-E Plus was designed and commercialized with an air sampling head. It can continuously measure and characterize bioaerosols with 0.3–100 µm size and has been reported to be the only instrument with integrated intelligence through graphics processing unit acceleration. Although discrimination between bioaerosols (bacteria and spores) and other contaminants was demonstrated at 99% precision by the producer, no peer-reviewed publications are yet available on the use of Rapid-E Plus for detecting bioaerosols in either laboratory settings or field studies.

For detecting and quantifying PBAP concentrations and calculating emission factors, the UV-APS and WIBS have both been applied in vastly contrasting field settings including high-altitude mountains, coastal regions, rural areas, densely-populated suburban and urban environments, megacities, indoor sites (such as hospitals, theatres, residences, and flood-affected housing), and industrial settings (such as wastewater treatment plants and swine containment buildings) [49,[83], [84], [85], [86], [87], [88]. SIBS has been widely used to detect the percentage of fluorescent particles in agricultural farms, dairy farms, urban background sites, wastewater treatment plants, and green waste composting facilities [89,90]. It has also been used in combination with other biochemical detection methods (matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) and qPCR) to identify constituents or metabolites in bioaerosol from compost, providing baseline training data for characterizing bioaerosols in real-world scenarios [91]. MBS has been used to identify and count PBAP within sea spray aerosol, as well as in a large multi-functional building [92,93]. For the PA-300, Crouzy et al. found that the system recorded 25 times higher concentrations of pollen than the manual approach [94]; for the Rapid-E, Daunys et al. analyzed the scattering and fluorescence data of 29 types of pollen and spores in the laboratory with a cluster approach and made natural pollen grouping possible [95]. Smit et al. measured seven airborne pollen types concentrations at both a 1-h and sub-hourly resolution [96].

In summary, the appearance of new instruments that have more fluorescence detection channels has been helpful for further discriminating between different bioaerosols, as well as between biological and non-biological compounds. Despite these advances, no real-time PBAP analysis technique can comprehensively detect all classes of bioaerosols, and biological aerosol classification methods are lacking. Given that the larger data sets produced by new LIF instruments greatly increase the load of data collection, processing, and calculation, it is imperative to improve numerical methods and tools for bioaerosol classification by using techniques such as hierarchical agglomerative cluster analysis (HACA), principal component analysis (PCA), linear discriminant analysis (LDA), and machine learning for discriminating bioaerosols from interferants. Understanding the differences between different versions of UV-LIF instruments is important for future applications as well as database comparisons. Finally, a comprehensive library of fluorescence spectra that can be reproduced and expanded for inter-laboratory comparisons and field observations needs to be built.

4. Raman spectroscopy

Raman spectroscopy is another rapid, specific, reliable, and non-invasive technique for real-time monitoring and diagnosis of bioaerosols that employs a laser to excite vibrational modes [97]. Normal Raman spectroscopy suffers from inherently weak Raman signals with only approximately 1 in every 10⁶–10⁸ photons being inelastically scattered. So far, a variety of enhanced Raman techniques have been developed to counteract this, such as coherent anti-Stokes Raman spectroscopy (CARS), surface-enhanced Raman spectroscopy (SERS), resonance Raman spectroscopy, and UV-Raman spectroscopy [98]. Among them, SERS and CARS are the two most popular techniques for rapid bioaerosol identification, both of which can suppress the fluorescence contained in biomolecules and greatly increase the Raman signal [81].

4.1. Surface-enhanced Raman spectroscopy

SERS, first observed by Fleischmann in 1974 [99], was developed to rapidly identify aerosolized pollen and bacteria (Table 3). Compared to normal Raman spectroscopy, the intensity using SERS is increased by more than 10⁶ to 10¹¹ fold using suspended silver, gold, or copper nanoparticles or a discontinuous metal surface. Sengupta et al. characterized airborne bacteria and pollen using SERS with an enhanced Raman spectrum by developing an impactor/collector system for sampling bioaerosols and either transferring them into a cuvette containing a suspension of nanocolloidal silver particles or directly transferring them to a circulating stream of the silver colloid using a micropump [100,101]. Schwarzmeier et al. combined a wet cyclone aerosol sampler (Coriolis µ) with an antibody-based microarray readout system for airborne E. coli analysis based on label-free SERS; they found that the entire detection process took less than 3 h and the selectivity was higher than that of antibody-antigen interaction (detection limit of 144 particles/cm³) [102]. Tahir et al. found that using the Andersen sampler and a commercial SERS substrate, Klarite (gold-coated inverted pyramid nanostructures), a strong SERS signal for live airborne E. coli was obtained. SERS-based Klarite was also used to identify changes in the E. coli cell membrane at different growth phases [103]. As shown in Fig. 5, Choi et al. developed a micro-optofluidic SERS platform, which ensures the collection of airborne particles by inertia and drag forces and simultaneously mixes them with SERS-active nanoparticles in a single microfluidic channel; using this platform, the bacterial aerosol detection limit reached 10² CFU/mL. This is also the first report of real-time airborne microorganism sampling and SERS analysis [104].

Table 3.

Bioaerosol components detected by SERS.

Bioaerosol samples	Laser radiation (nm)	SERS surface coating	Observation	References
E. coli	514	Ag colloid	Aerosolized bacterial spectra are similar to those of stock suspensions	[100]
E. coli	532	Ag nanoparticles	The first platform to realize the sampling and SERS real-time analysis of bioaerosol, with a detection limit of about 100 CFU/mL	[104]
E. coli	785	Klarite	Estimate the percentage of live E. coli in the air	[103]
E. coli	633	Ag colloid	For quantitative analysis, the detection limit is 144 particles per cubic centimeter	[97]
Pseudomonas aeruginosa, Salmonella typhimurium	514	Ag colloid	Aerosolized bacterial spectra are similar to those of stock suspensions	[100]
Pollen [cottonwood], Pollen [redwood]	514	Ag colloid	Compared with bacteria, the influence of background water vibration on pollen is less.	[100]
Staphylococcus epidermidis, Mycobacterium luteum, Bacillus subtilis	532	Ag nanoparticles	The first platform to realize the sampling and SERS real-time analysis of bioaerosol, with a detection limit of about 100 CFU/mL	[104]

Fig. 5.

Design of a continuous optofluidic SERS system for detecting bioaerosols (from ref.[104]).

4.2. Coherent anti-Stokes Raman spectroscopy

CARS is a nonlinear optical process based on four-wave mixing signals that increase the effect of Raman scattering by orders of magnitude through resonance enhancement [98]. Since Zumbusch et al. first pioneered CARS for live cell imaging [105], its application in bioaerosol detection has been continually reported (Table 4). Ooi et al. combined the Lorentz-Mie theory with femtosecond dynamics in a nonlinear process to detect about 10⁷ photons from spores placed 1 km behind the detector; this suggests that this technique has the potential to remotely identify airborne particles, particularly pathogenic spores [106]. Petrov et al. compared CARS with spontaneous Raman microspectroscopy, and found that it was more efficient in detecting B. subtilis endospores, with the CARS signal being 100 times stronger than that from Raman microspectroscopy [107]. Deckert et al. developed a viral detection system combining tip-enhanced Raman technology and femtosecond adaptive spectroscopic techniques (FAST) CARS, which was able to map individual viral particles with nanometer resolution and high sensitivity [108]; in particular, this report provides hope for the detection and identification of single viral particles.

Table 4.

Bioaerosol components detected by CARS.

Bioaerosol samples	CARS model	Pulse properties	Observation	References
Bacterial spores	FAST CARS	Ultra short ∼50 fs	The sensitivity of dipicolinic acid (DPA) is 106 times higher than that of non-resonant Raman	[109]
Bacillus anthracis	Theory of femtosecond coherent anti-Stokes Raman backscattering enhanced by quantum coherence	Ultra short ∼50 fs	The backscattered photons are sufficient to construct accurate spectral data	[106]
Bacillus subtilis	CARS	Broadband continuous wave	CARS is generally 100 times more efficient than conventional Raman	[107]
H1N1, CVB3	Femtosecond adaptive spectroscopic techniques with enhanced resolution via coherent anti-Stokes Raman scattering	Ultra short ∼50 fs	Map a single virus particle with nanometer resolution and chemical specificity.	[108]

In summary, both SERS and CARS can enhance the signal intensity and characterize the Raman peaks more effectively, and can be used to detect airborne bacteria and their chemical composition, endospores, and pollen in the laboratory; however, detection has not been accomplished in field experiments [109]. SERS has a short assay time and requires far smaller concentrations of analytes than traditional Raman spectroscopy; BWA surrogates and kinetic processes of endospore germination can also be detected by SERS [110,111]. Using Klarite as a substrate, the changes in the bacterial envelope during different cell growth phases and the number of single live and dead bacteria can also be revealed [103]. CARS can be used for standoff detection of biosamples (such as bacterial endospores) more efficiently than spontaneous Raman spectroscopy and remote sensing by suppressing the non-resonant background. Using FAST for CARS, a single virus particle with nanometer resolution and chemical specificity can directly be identified.

5. Mass spectrometry

Mass spectrometers can also be used for the online detection of bioaerosols. Based on the presence or absence of a matrix, this technique can be divided into two types. The first is a matrix-free laser desorption/ionization method, with BAMS as the main representative. This is a real-time single particle analytical method with the ability to sample and analyze cells of sizes from 0.5 to 7 µm that eliminates the need for reagent and sample preparation, making detection and analysis of M. tuberculosis aerosols more accurate [112,113]; however, the detection limit and sensitivity are relatively low due to the low ion transmission efficiency [114]. To overcome this limitation, laser desorption/ionization has been gradually replaced by MALDI. This has led to the development of the second representative mass spectrometry (MS) technology for bioaerosol analysis, MALDI-TOF MS. Here, a matrix is added online to coat the bioaerosols in an evaporation/condensation tank in front of the TOF MS [115]. This provides a soft ionization technique for large bioaerosols with a minimum amount of fragmentation. As compared with BAMS, MALDI-TOF MS has been employed for the online detection of very similar particle types [116]. Its application in detecting pollen lipids [117] and bacterial and fungal isolates has been recently reported [118].

5.1. Bioaerosol mass spectrometry

BAMS, developed at the Lawrence Livermore National Laboratory, USA, is a spectroscopic method with high specificity and sensitivity. It allows the analysis of bioaerosols in real time and permits discrimination between different species of bacteria [113,119]. When aerosol particles are sampled into the inlet of a BAMS system, they are accelerated into a vacuum system through a nozzle; cells are then sized using the time delay measured between two scattering lasers. Next, a 266 nm laser is used to desorb and ionize the molecules in the particles, before they are analyzed by dual polarity TOF-MS. The analysis process takes far less than a second per particle with little to no sample preparation required [113,120]. As early as 1985, Sinha et al. identified B. subtilis, B. cereus, and Pseudomonas putida by particle beam MS. Specifically, the particle beam of the cell was generated and then volatilized and ionized in the ion source of a mass spectrometer; finally, the average intensities of different mass peaks were measured to generate the mass spectrum [121]. Additionally, Fergenson et al. combined single-particle laser desorption/ionization TOF-MS with real-time pattern recognition for the real-time and reagent-free characterization of bioaerosols without sample pretreatment. The results showed that this method could distinguish two spores (B. thuringiensis and Bifidobacterium atrophum) very accurately and with a sensitivity of 92%; however, the observed low mass peak (mass to charge ratio (m/z) ≤ 200 Da) may be due to the ionization and mass spectrometer limitation [114]. Srivastava et al. used BAMS to fully characterize the mass spectral signatures of B. atrophaeus spores, and found that m/z = +74 was the marker peak for distinctions between B. atrophaeus and B. thuringiensis spores; however, this marker must be used with caution as it depends on the growth media used [112]. Tobias et al. applied BAMS to rapidly detect airborne Mycobacterium smegmatis and M. tuberculosis H37Ra in real-time using a marker peak at m/z = −421; its accuracy and sensitivity were quantitatively established for the first time using APS. The detection limit of this method is still relatively low, which may be attributed to the low sampling efficiency of particles in the small inlet [120]. Frank et al. used BAMS to detect Mycobacterium spp. embedded in lung surfactants, which helped demonstrate that BAMS could be applied for the real-time diagnosis of tuberculosis [113].

5.2. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry

The 266 nm laser power used by BAMS for laser desorption/ionization is not suitable for biomolecules with molecular weights > 1000 Da due to the production of abundant low-mass ions fragmentation during ionization [112]. Karas and Hillenkamp developed MALDI-TOF MS in 1988 to produce complete protonated biomolecules by providing a gentler ionization method [122]. For MALDI-TOF MS, a matrix is used to coat the bioaerosols in real-time (Fig. 6) [123]. For example, Stowers et al. coupled the inlet of the aerosol TOF MS with a flow cell containing the matrix; when the TOF MS sampled bioaerosols through the flow cell, it coated them with a matrix, generating MALDI mass spectra. By investigating the mass spectra generated from airborne erythromycin, gramicidin-S, or B. subtilis spores using picolinic acid, 3-nitrobenzyl alcohol, and sinapinic acid as matrices, the research group found that MALDI-TOF MS was able to identify bioaerosols nearly in real-time, with less than a 5 s residence time; however, the lower field strengths in the current ion source resulted in a lower mass range [115]. Wuijckhuijse et al. combined aerosol TOF MS (ATOFMS) with laser-induced fluorescence and MALDI MS to investigate the effect of laser power and sample pretreatment on the mass spectra quality of B. subtilis spores. The group found that a narrow range of lasers provided the optimal signals; the online matrix (ferulic acid) application gave high-quality mass spectra, and the width of the particle beam affected ion flight time and decreased mass accuracy and resolution [124]. Bogan et al. created polyethylene glycol-containing particles that produced an array of spectra that were reproducible and easy to recognize. These particles can be used for BMS calibration for measuring complex samples and for developing online aerosol laser desorption ionization mass spectrometer platforms [125]. McJimpsey et al. found that the matrix, the ratio of matrix-to-analyte, and the diameter of the particle affect efficient ion generation in aerosol MALDI MS and that the analyte ion signal changed linearly with the absorbed energy, volume, and surface area of the particles [116]. Liang et al. used MALDI-TOF lipid profiling to detect the pollen grains and obtained more mass peaks and higher signal intensity when using microwave-assisted formic acid digestion, which avoided preparation of the analyte/matrix layer achieving continuous sampling and detection [117]. Osa et al. also compared the concordance rates between MALDI-TOF MS and traditional methods and found that the rate of the former was 99.9% (genus) and 96.2% (species). Furthermore, 3,502 of 3,530 bacterial and fungal isolates were identified by MALDI-TOF MS, suggesting that it could be used as a rapid diagnostic method with high accuracy for microorganisms [126].

Fig. 6.

Schematic representation of a direct in situ MALDI TOF MS system for the rapid bioaerosols detection (from ref.[123]).

In summary, for real-time bioaerosol detection, BAMS focuses on accuracy, while MALDI-TOF MS focuses on universality. Both methods can be used to detect and distinguish airborne bacteria and fungi in a laboratory chamber, but not in the field. BAMS has higher accuracy but lower sensitivity for high-mass biomolecules [127]. Increasing the particle inlet and removing the limitation of the ionization process may improve the overall detection limit, but higher and more specific m/z biomarkers still need to be identified. MALDI-TOF MS requires the least number of substrates and reagents, and could provide high sensitivity and a higher operational mass range [116] that could be used to identify pathogens; however, its accuracy needs to be improved. A database that contains reference-quality fingerprints of single and mixed species is still required, and the reference spectrum of MALDI ATOFMS still needs to be enriched [123].

6. Conclusion and future outlook

The recent outbreak of infectious diseases, such as COVID-19 caused by the SARS-CoV-2 virus, has been a large impetus to investigate bioaerosol detection. Providing simple, specific, and affordable real-time detection methods is important for bioaerosol surveillance network construction. This review sheds light on the strengths and weaknesses of several types of online bioaerosol detection techniques, along with their working principles, sensitivity, efficacy, and application in the laboratory and some field studies (Table 5). The following key conclusions are highlighted:

Table 5.

Summary of the four types bioaerosol detecting techniques.

Bioaerosol detecting techniques	Principle	Application	Advantages	Disadvantages
ATP bioluminescence measurement	The amount of fluorescence produced during ATP's reaction is proportional to ATP concentration contained in the biomass	Viable airborne bacteria and fungi detection	Quantitative analysis; simple, cost-effective and rapid	Not suitable for field test; cannot differentiate microorganisms on the species level
UV-LIF spectroscopy	Biofluorophores in the microorganisms can auto-fluoresce under UV light/laser exposure	Airborne bacteria, fungi and pollen detection	Quantitative and sometimes qualitative analysis, non-invasive with high time resolution	Difficult for virus detection; will affect by non-biological compounds with fluorescence
Raman spectroscopy	Based on the Raman effect; by capturing the intrinsic molecular vibrations in the samples excited by a laser	Airborne bacteria and their chemical composition, endospores and pollen detection	Quantitative and qualitative analysis; rapid, specific, reliable and non-invasive	Difficult for online detection; not suitable for field test; expensive, can identify virus
Mass spectrometry	Based on ions generated by laser desorption or pyrolysis	Airborne bacteria and fungi	Quantitative and qualitative analysis; discriminating microorganisms with high specificity and sensitivity	Difficult for online detection; not suitable for field test; expensive

(1) The ATP bioluminescence assay has been combined with different modified bioaerosol sampling techniques for rapid and real-time/near real-time bioaerosol monitoring with simple, cost-effective, and rapid characteristics. However, this method only presents the concentration of all living bacteria, fungi, and other cells in the bioaerosol and cannot differentiate or specify pathogens or microorganisms at the species level.

(2) UV-LIF instruments allow non-invasive bioaerosol detection with fast resolution. UV-APS and WIBS have been commonly used for evaluating the size, total fluorescence intensity, and concentration of bioaerosols both in laboratory and field studies. The appearance of BioScout, SIBS, and Rapid-E with more channels has increased the sensitivity and improved the prospects of bioaerosol monitoring, but they still cannot comprehensively detect all classes of bioaerosols; the distinction between weakly fluorescent, interfering abiotic particles and weakly fluorescing bioaerosols is also still very challenging. Improved numerical methods as well as tools and machine learning are urgently required to discriminate bioaerosols and interferants, and a comprehensive library of fluorescence spectra needs to be constructed.

(3) Raman spectroscopy is a rapid, specific, reliable, and non-invasive technology for detecting airborne bacteria and their chemical composition, endospores, and pollen in the laboratory. For SERS, the addition of nanocolloidal silver, copper, or gold particles to the analyte greatly enhances Raman scattering by 10⁶ to 10¹¹ fold. This technique also involves a short assay time and requires a far smaller concentration of analytes. CARS can be used to detect biosamples at far distances and identify viruses with nanometer resolution and chemical specificity. Future studies need to couple with other analytical tools to improve the selectivity, sensitivity, and real-time analysis of Raman spectroscopy.

(4) Mass spectrometers can be used for analyzing bioaerosols in real time and for discriminating different bacterial species with high specificity and sensitivity. BAMS has higher accuracy and eliminates the need for reagent and sample preparation. Its relatively low detection limit and sensitivity could be improved by increasing the particle inlet and removing the limitation of the ionization process. MALDI-TOF MS can be used for the online detection of pollen lipids, and bacterial and fungal isolates with higher sensitivity; however, accuracy improvement is needed and a database that contains reference-quality fingerprints as a reference spectrum also needs to be built in the near future.

In summary, all existing modern real-time bioaerosol detection methods are moving quickly towards the development of more specific, efficient, reproducible, easy-to-use, and easily accessed techniques. Reaching these goals is still challenging due to the complexity and variability of bioaerosols, the interference of abiotic components, as well as the influence of environmental factors. Most current technologies are not suitable for field applications, and can only be used for quantitative analysis. For affordable ATP bioluminescence assays, highly efficient bioaerosol capture techniques combined with conventional detection methods such as qPCR are favorable for identifying specific bacterial/fungal species. For inexpensive and non-invasive LIF instruments, developing cheaper dual UV LED-based bioaerosol monitors and comparing their performance with that of more expensive models as mentioned above is urgently needed. For the costly Raman spectroscopy and BAMS techniques, future work needs to focus on more laboratory and field validation tests for bioaerosol detection, constructing (fluorescence) spectra libraries and databases for elucidating spectrally integrated signals, and improving measurement selectivity of bioaerosols.

Author contributions

Taicheng An: conceptualization, supervision, writing – review & editing. Zhishu Liang performed the research, analyzed the data and wrote-original draft. Zhen Chen researching and writing of the paper. Guiying Li: writing – review & editing.

Declaration of competing interest

The authors declare that they have no conflicts of interest in this work.

Biography

Taicheng An(BRID: 08993.00.08792) received his PhD from Sun Yat-sen University. He is currently the dean of School of Environmental Sciences and Engineering, Guangdong University of Technology. His research interests mainly focus on environmental geochemistry and health. He has published more than 540 SCI papers in Nat. Commun., PNAS, JACS, etc. He is one of the most cited Chinese authors in environmental sciences by Elsevier's Scopus database from 2014 to 2022, the winner of NSFC Distinguished Young Scholars, and Distinguished Professor of Chang Jiang Scholars. He also served as associated editors of Appl. Catal. B: Environ., and Crit. Rev. ES&T .