1. Introduction

A method to achieve robust, real-time aerosol detection or classification will rely on rapidly interrogating individual aerosol particles. Optically based, spectroscopic interrogation techniques are well suited for in situ implementation due to fast response times and the ability to perform on-the-fly or non-contact measurements. Established optical methods that have been previously explored for single particle detection are elastic and Raman scattering of light, fluorescent emission spectroscopy and absorption spectroscopy. Of these, fluorescent emission and Raman scattering have potential to identify the composition of single particles using limited collection angles and without prior knowledge of the particle shape. The spectroscopic signatures from data acquired from Raman scattering can be very specific, but are also weak, while fluorescence signatures are less specific but typically 2 to 3 orders of magnitude stronger.

The requirements for a real-time, single-particle analysis would be to characterize picogram quantities of material within millisecond time periods. Biological materials, such as microorganisms, typically include fluorophores, such as tryptophan, riboflavin, and NADH [1,2]. Consequently, micron-sized ambient biological aerosol particles commonly contain sufficient fluorescent material for detection using the ultraviolet (UV) spectral region. This has led to development of single particle on-the-fly fluorescence sensors for biological aerosol detection systems that can measure fluorescence and elastic scattering signatures from single cell bacterium and fungal cells [3–7]. Currently, the detection limits of such sensors are typically in a range from 0.7 µm to 2 µm sized particles, depending on their specific detection configuration. Increased fluorescence signal sensitivity would permit shorter interrogation times and increased particle throughput while also providing an improved ability to perform multiplexing from multiple wavelength excitations. Additionally, increased sensitivity would result in smaller particle size detection limits, which may be important for detection of smaller organisms such as viruses and various environmental pollutants in real-time.

Improving fluorescence detection capability is attractive, but the ability to perform single particle Raman detection on-the-fly, would address an existing capability gap for detection of materials that are non-fluorescent. Due to low signal-to-noise limitations however, Raman studies of aerosols have thus far been performed only for ensembles of flowing aerosol particles or for single particles levitated and held in a trap [8–11]. Extrapolations from Raman spectra studies and reported Raman cross-sections for chemicals of interest [12] suggest that an enhancement of about 10³ would be required to achieve real-time Raman detection of micron-sized aerosols.

A commonly used method to increase either Raman or emission cross-sections exploits the enhanced fields created by metallic nanoparticles (MNPs). Materials in the presence of MNPs have reported enhancements of fluorescence, termed Metal Enhanced Fluorescence (MEF) [13] as well as enhancement in Raman spectra termed Surface Enhanced Raman Spectroscopy (SERS) with enhancement factors ranging from 10³ to 10¹⁴ [14,15]. Various configurations have been explored over the past several decades that place target materials onto specialized MNP substrates or held in suspension with MNP colloids. These techniques have been previously explored for enhancing Raman and fluorescence signatures in biological and chemical defense applications by several authors [16–20], including collected biological aerosols in liquid [17], but thus far has not been applied directly to individual aerosols in situ.

2. Experimental approach

In this paper we generate micron-sized aerosol particles together with commercially available MNPs and acquire their fluorescence spectra by utilizing a single-particle, on-the-fly, two wavelength fluorescence excitation system [7,21]. To our knowledge this approach has not been previously reported. Aerosol particles in this analysis are prepared with MNPs located either on the surface of a solid spherical bead or embedded within a protein matrix. The method uses droplets generated from liquid suspensions containing known concentrations of the materials of interest as well as MNPs. Droplets containing the mixtures were produced using a piezo-electric droplet generator (MicroFab Technologies Inc.) and then passed through a heated, drying column with flowing dry nitrogen as the carrier gas, to form the final dry aerosol particles [22]. Particles with surface coatings of MNPs were created using fluorescent polystyrene latex spheres (PSL) suspended in water with MNPs, whereas particles embedded with MNPs were created from a solution of proteins and suspension of MNPs in water. For the surface coated particles, 2 µm diameter, 450 nm dye-doped PSLs (Thermo Scientific part #B0200) were chosen. When producing these aerosol particles, the concentration of PSLs was adjusted, such that predominantly no more than one PSL would be contained in each droplet. For MNPs embedded aerosols, an amino acid, DL-tryptophan (Sigma-Aldrich part#16269-8) and protein, ovalbumin (Sigma-Aldrich part #A2512-1G) were used as a matrix material, and the dissolved concentration was regulated to produce consistent final particle diameters of about 2.5 µm and 2.2 µm respectively. Aerodynamic particle diameters of the generated aerosols were measured with an Aerosol Particle Sizer, APS (TSI Inc., Model #APS 3321). The MNPs used in this study are commercially available from Ted Pella, Inc., as citrate stabilized gold/silver colloid suspensions in water. To prevent aggregation of the MNPs, the prepared suspensions of core particle material and MNPs was sonicated prior to droplet dissemination.

3. Generation of aerosols and fluorescence measurements

Figure 1 shows typical SEM images of two example cases of generated aerosols: (a) a 2 µm PSL particle coated with 2 × 10⁴, 10 nm gold MNPs on the surface resulting in a 25% surface coverage with MNPs and (b) a 2.5 µm tryptophan particle with 3 × 10⁴, 10 nm gold MNPs embedded throughout the particle resulting in a 0.2% volume fraction coverage with MNPs. For this image, the generated aerosol particles were collected by impaction onto a sample substrate located in the outlet air flow from the drying column. The SEM images in Fig. 1 show that the aerosol particles are either coated or interspersed with MNPs. The MNPs have partially aggregated, especially in the case of the surface coated PSL due to higher number density of MNPs on the surface. MNP aggregation could have occurred either in suspension (even though, noticeable color change was not observed) or (more likely) as the generated droplet evaporated and dried. The colloidal MNPs used in this study come suspended as citrate stabilized with a net negative surface charge that keeps the MNPs stable in water suspension. Several factors such as the chemical environment of the suspension and the mechanism of droplet production (which typically imparts a charge to the drop) could result in neutralization of the surface charge of the MNPs, and thus allow aggregation. The contributing factors leading to aggregation will need to be investigated and possibly controlled in a future study.

 

Fig. 1 SEM images of (a) 2 µm PSL particle coated with 2 × 10⁴, 10 nm gold MNPs on the surface resulting in a 25% surface coverage and (b) 2.5 µm tryptophan particle with 3 × 10⁴, 10 nm gold MNPs embedded throughout the particle resulting in a 0.2% volume fraction coverage.

Download Full Size | PDF

As mentioned earlier, fluorescence data from the generated aerosols was acquired using our Aerosol Interrogation Module (AIM) system, a two-wavelength laser excitation, single-particle aerosol analyzer that was developed previously [21]. The AIM device illuminates and characterizes individual aerosol particles in the 0.7 to 10 μm size range using elastic scatter from a 785 nm CW laser, and acquires fluorescence intensities excited by two pulsed UV lasers at 266 nm and 355 nm respectively. The observation of 785 nm elastic scatter from individual particles acts to sequentially trigger single pulses from each of the two UV excitation lasers, and also provides a relative indication of the particle size from the differential scattering cross-section. Fluorescence emission from each particle is collected in four broad spectral bands centered at 324 nm (303 to 344 nm), 390 nm (380 to 400 nm), 453 (408 to 497 nm) and 553 nm (502 to 603 nm) for the 266 nm excitation and in the latter three bands for the 355 nm excitation.

Aerosols composed of either tryptophan, ovalbumin or dye-doped PSLs were generated and the fluorescence of the individual aerosols was recorded. Fluorescence intensities for all four emission bands were recorded and summed for each particle from the 266 nm excitation and similarly for the three bands available from the 355 nm excitation. Experimental results are plotted in Fig. 2 using logarithmic scales for the two summed emission intensity values, with each dot representing the fluorescence from a single aerosol particle. The same particle material concentrations of these samples were then prepared with the addition of suspended 10 nm gold MNPs (to yield concentrations of approximately 900 MNPs in each 60 to 80 µm droplet), and droplets were generated and dried to form aerosol particles in the same way. Fluorescence data from the three aerosol sample materials with the added MNPs are also plotted in Fig. 2. The differences observed in the fluorescence of the aerosol populations indicate effects due to the presence of MNPs, which can be either positive (enhancement) or negative (quenching) and the details of these variations are discussed in Sec. 5. In some cases the differences appear slight due to the logarithmic scales, but for example, the tryptophan sample with MNPs in Fig. 2 shows a significant (2x) enhancement over the same sample without MNPs.

 

Fig. 2 Total absolute fluorescence intensities from aerosol particles of tryptophan, ovalbumin and dye-doped PSLs are plotted for excitations at 266 nm and 355 nm. Samples were measured with, and without, the addition of approximately 900, 10 nm gold MNPs. Each point on the plot corresponds to the fluorescence emission from a single aerosol particle. The 785 nm elastic scattering signal, which is indicative of the size of the aerosols, is shown in the inset illustrating that the addition of MNPs did not significantly alter the size distribution of the resultant aerosols.

Download Full Size | PDF

For each of these three aerosol particle materials, the operating parameters of the micro-droplet generator required slight adjustments to optimize stable droplet production [22], and care was taken to ensure that the same size droplets were generated from each of these materials both with and without the MNPs. Ultimately a standard deviation of 6% was achieved for the droplet diameter using each material, regardless of the presence of MNPs. The size of the generated dried aerosols can be determined by the APS, and the relative size can be inferred from the elastic scattering intensity from the 785 nm CW laser. The 785 nm laser data was used to verify the size of each particle as they were simultaneously interrogated for fluorescence, and the resultant histogram distribution is plotted as an inset in Fig. 2 using the same color scheme as the fluorescence data. The shapes of these particle size distributions for each particle sample types are identical regardless of the presence or absence of MNPs indicating that the size of the generated aerosols was held effectively constant during these measurements. Therefore observed variations in the fluorescence spectra are indicative of a change in the spectra and not due to variation in aerosol size. The secondary peak observed in the inset of Fig. 2, for PSL samples (red and orange traces) represents the aerosols generated from the smaller population of droplets containing two or more PSLs and are responsible for the outlying fluorescence data group located to the top-right of Fig. 2. As mentioned earlier, the concentration was chosen so that predominantly only one PSL will be present in each droplet, however there was a 26% probability of generating droplets with more than one PSL [22].

The mean absolute fluorescence measured in each of (a) the four bands for the 266 nm excitation and (b) the three bands for the 355 nm excitations for the three samples, tryptophan, ovalbumin and 2 µm dye-doped PSL; without the addition of MNPs, are plotted as bar graphs in Fig. 3. These plots are computed from the same data set shown in Fig. 2, and the error bars represent the standard deviation of the measured fluorescence signal for 10,000 individual particles for each aerosol population. As expected, the predominant fluorescence for amino acids and proteins are apparent in the UV bands excited by 266 nm excitation, as seen in Fig. 3(a) and comparatively low fluorescence intensities excited by 355 nm as seen in Fig. 3(b). For the dye-doped PSL samples, fluorescence is generated by their dye component primarily in the two visible bands spanning the 400 to 600 nm spectral region, as well as from the intrinsic polymer, which is observed in the UV bands spanning from 300 to 400 nm.

 

Fig. 3 Absolute fluorescence emission is shown for the three target aerosols; tryptophan, ovalbumin and 2 µm dye-doped PSL, measured in (a) four bands using 266 nm excitation and (b) three bands using 355 nm excitation. The measurements were performed on populations of approximately 10,000 aerosols for each sample, with the mean and the standard deviations shown.

Download Full Size | PDF

4. Fluorescence measurement as a function of MNP concentration

In the second part of the experiment, we varied the concentration of the MNPs (number of MNPs per aerosol) in a controlled manner and observed the resultant effect on the measured fluorescence emission intensity from the three target materials. To acquire aerosol images, 2 µm PSLs were suspended with varying concentrations of 10 nm gold MNPs yielding single PSL aerosol particles coated with none, 3 x 10³ and 10 x 10³ MNPs, corresponding to surface coverage of 0, 4 and 13 percent, respectively. Representative SEM images of collected samples from each of these aerosolized PSL populations are shown in Fig. 4. The surface ofthe PSL with no MNPs shows a clean surface while the SEM images of PSLs with varying higher concentrations of MNPs shows progressively denser coverage, confirming that adding higher concentrations of MNPs to the suspension does result in higher concentration of MNPs coating on the PSLs. Partial agglomeration of MNPs noted in the SEM images in Fig. 1 is also observed in Fig. 4 at both concentrations.

 

Fig. 4 SEM images of 2 μm PSLs with addition of 0, 3x10³ and 10x10³, 10 nm gold MNPs on the surface of the particle corresponding to surface coverage of 0, 4% and 13%. The SEM images confirm the deposition of higher density of MNPs coating the PSL as the concentration of the MNPs was increased in the solution to generate the droplets. The deposition also shows agglomeration of the MNPs.

Download Full Size | PDF

We then measured fluorescence from the aerosols of the three target materials; PSL, ovalbumin, and tryptophan with varying concentrations of MNPs. The concentration of the MNPs was varied such that the number of MNPs per aerosol ranged over three orders of magnitude, approximately from 10¹ to 10⁴. Using the highest concentrations of MNPs, we expect a volume fraction of MNPs to be 0.4% for ovalbumin and 0.08% for tryptophan aerosols, where the MNPs are embedded throughout the aerosol particle and a surface coverage of 20% for the 2 µm dye-doped PSLs. The fluorescence was measured for both the 266 nm and 355 nm excitations in the same bands and summed, as done in the preceding measurements. The results were then normalized to the total in-band fluorescence measured in the absence of MNPs and plotted as a function of the MNP concentration in Fig. 5 with the error bars representing the standard deviation from populations of 10,000 aerosol particles per sample. Figure 5(a) shows that the 266 nm excited fluorescence shows minimal change andexhibits quenching for increased concentrations of MNPs for all three target materials. However, different behaviors are observed for the 355 nm excitation in Fig. 5(b). Ovalbumin particles show fluorescence enhancement as a function of increasing concentration of MNPs approaching a factor of three for the highest concentration used in this study. The fluorescence of tryptophan particles initially shows enhancement as a function of concentration, nearing a factor of 2 before quenching as the concentration is further increased. Data from the dye-doped PSL particles shows only quenching of the fluorescence for increased MNP concentration.

 

Fig. 5 Scaled fluorescence intensity of tryptophan, ovalbumin and 2 µm dye-doped PSL particles are plotted as a function of concentration of the addition of gold MNPs. (a) The fluorescence signals measured in the four emission bands using 266 nm laser excitation has been summed and normalized to the fluorescence measured in the absence of MNPs, and similarly (b) the sum of the fluorescence in the three emission bands using 355 nm laser excitation has been summed and normalized to the fluorescence measured in the absence of MNPs

Download Full Size | PDF

5. Results and discussion

Greater insight can be gained by examining the fluorescence measured in the separate bands for each excitation laser wavelength. Intensities measured for the three samples in the four emission bands from the 266 nm excitation and three emission bands from the 355 nm excitation are plotted in Figs. 6(a)–6(c) and Figs. 6(d)–6(f), respectively. The fluorescence data in all the bands are normalized to the emission intensities measured from aerosols generated without MNPs for each respective band. Even though the 266 nm excited total fluorescence (sum of the fluorescence from all bands) shown in Fig. 5(a) are quenched with increasing concentrations of MNPs, it is notable that some enhancement occurs for both tryptophan and ovalbumin in the weaker, long-wavelength emission bands as can be seen in Figs. 6(a) and 6(b). This was masked when looking only at the emission summed over all channels, as the predominant fluorescence emission occurs in the UV bands for biological materials like tryptophan and ovalbumin as discussed in Fig. 3. However, for dye-doped PSL samples, quenching is exhibited in all bands, for both excitations as shown in Figs. 6(c) and 6(f). For the 355 nm excitation profiles, similar correlations are evident with that of 266 nm excitation, except now tryptophan in Fig. 6(d) and ovalbumin in Fig. 6(e) show enhancements even in the 390 nm UV band. Additionally, tryptophan shows higher enhancement in the UV emission band compared to the Vis emission band, Fig. 6(d).

 

Fig. 6 Scaled fluorescence intensities are plotted as a function of concentration of the addition of gold MNPs. Plots of fluorescence emission measured using 266 nm excitation are shown in (a), (b) and (c), and those with 355 nm excitation are shown in (d), (e) and (f) for tryptophan, ovalbumin and 2 µm 450 nm dye-doped PSL aerosols respectively. The fluorescence measurements have been normalized to the emission measured for the aerosols without the addition of MNPs in the respective bands.

Download Full Size | PDF

Although absorption spectra of the individual aerosols were not measured, it can be extrapolated from prior work of Creighton and Eadon [23]. The absorption cross-section of 10 nm gold MNPs is broad through the UV-Vis range with a plasmon resonance peak near 520 nm. They report the spectra increasing in strength and shifting to longer wavelengths when the MNPs are surrounded by water as compared to being in a vacuum [23]. In our case, with tryptophan and ovalbumin, the MNPs are embedded throughout the aerosols, whose index of refraction is close to water (1.33 < n < 1.4), so that the known absorption cross-section of MNP in water with spectral peak at 520 nm is a reasonable assumption. In the case of aerosols composed of PSL, the MNPs are on the surface of the aerosol and the weaker (factor of 3.2 lower) absorption curve corresponding to MNP in vacuum should be more applicable.

There are a number of possible phenomena at play here. The fluorescent emission of a molecule can be affected by the presence of MNPs by different mechanisms. Two principal effects, which are competing processes, are due either to field enhancements or to the non-radiative damping of the fluorescence emission owing to energy transfer from the fluorescing molecule to the MNPs [24,25]. The former can result in an overall fluorescence enhancement either due to enhancement of the excitation wavelength or the emitted light, while the latter creates an overall decrease in fluorescent signal. Both of these effects may be apparent in the experimental results illustrated in Figs. 5 and 6. For distances between an MNP and a fluorescent molecule that are small (< 5 nm) quenching is the dominant process, while enhancement is both predicted and observed for larger separations [26]. These predicted behaviors have also been experimentally demonstrated by Liu et al. [27] and Anger et al. [26]. The enhancement and quenching of tryptophan fluorescence observed for the 355 nmexcitation as a function of the concentration of MNPs in Fig. 5(b) has been observed and reported by others [28–31] (including solutions of tryptophan with silver colloids in water by Caires et al. [29]). However, in a multiple fluorophore and MNP composition, as in the case of aerosols embedded with MNPs or liquid colloids, it is not straightforward to model the rate of quenching or enhancement as a function of concentration for randomized systems.

Several authors have studied quenching of fluorescence by layers of metal nano-particles [32,33], and recently Zhang et al. [34] have reported such a study in quantum dots and MNPs. The authors have been able to experimentally observe and corroborate their results with nanometal surface energy transfer theory (NSET), which provides an explanation of the dependence of the wavelength, concentration and distance pertaining to the quenching of fluorescence from quantum dots due to gold MNPs [34]. They observed that the quenching efficiency is not only higher for emission wavelengths farther from the plasmon resonance peak, but is even a stronger function of the distance between the quantum dots and the MNPs (i.e., quenching occurs at greater distance for wavelengths farther from the plasmon resonance). This result is consistent with our observed increased quenching for the shorter wavelength UV bands compared to the longer Vis wavelength bands in Figs. 6(a) and 6(b) that have relatively less spectral overlap with the plasmon resonance peak.

The spectral position of the excitation in relation to the emission band or to the plasmon resonance band of the MNPs may also be contributing to the behavior of the fluorescence data in Figs. 5 and 6. Fluorescence enhancement will be proportional to the frequency overlap of the plasmon resonance of the MNPs with that of the excitation wavelength [19,35–37] and also the fluorescence emission of the target material [38]. The fluorescence enhancements at 355 nm excitation, Figs. 6(d) and 6(e) are higher compared to the enhancements at 266 nm, Figs. 6(a) and 6(b) for the same samples, this could be perhaps due to the relative higher coupling efficiency of the 355 nm excitation to the plasmon resonance band of the MNP at 520 nm. Comparing the coupling efficiency of the emission wavelength to the plasmon resonance of the MNP, however is not straightforward as the efficiency of quenching also comes into play. The fluorescence enhancement is observed to be higher for the longer wavelength emission bands for both excitations in ovalbumin, Figs. 6(b) and 6(e) and the 266 nm excited tryptophan, Fig. 6(a), which is in agreement with Tam et al. [38], in that increased fluorescence enhancement is observed for wavelengths that have greater overlap with the plasmon frequency. This correlation, however, is not supported by our data from tryptophan excited with 355 nm, Fig. 6(d), where the shorter wavelengths exhibit greater fluorescence enhancement. It is not clear why the fluorescence band centered at 390 nm show either quenching (266 nm excitation) or enhancement (355 nm excitation) as a function of the excitation wavelength for the same samples.

Differences in the experimental results between the amino acid and protein particle sample materials used in our study may also be interpreted in terms of their relative molecular sizes. Using the molar mass of tryptophan, 204.23 g, the number of molecules present in a 2.5 µm diameter spherical aerosol particle can be estimated to be about 10¹⁰, and assuming spherically shaped molecules, yields a radius of the amino acid of about 0.4 nm. Ovalbumin is a long chain protein, whose molecular weight is approximately 45kDa, which yields ~10⁸ protein molecules in a 2.2 µm aerosol particle. Again, assuming a spherical shape, results in a molecular radius of 2.4 nm. This difference of a factor of five in size between the amino acid and protein effectively translates to a much larger spacing for ovalbumin compared to tryptophan between a potentially enhancing MNP and the active molecular region contributing to fluorescence. The observed results shown in Fig. 6 as a function of MNP concentration are consistent with this consideration, since enhancement of the fluorescence occurs at a lower concentration of MNPs for the smaller tryptophan molecule compared to the ovalbumin molecules.

For the highest concentration of MNPs added (~10⁴) per aerosol particle, the ratio of MNP to tryptophan molecules is 1:10⁶; while for ovalbumin molecules, it is 1:10⁴. Using 10 nm gold MNPs and assuming a distance of 5 to 10 nm from the MNP for enhancement influence (estimation from [39,40]), we expect a contribution from less than 1.4 × 10⁴ tryptophan molecules. For measured enhancement of a factor of 2 in the overall fluorescence of tryptophan aerosol at MNP concentration of 10³ in Fig. 5(b) this would mean that the enhancement factor from each of the contributing molecules will be around 700. This enhancement factor drops to 80 if the distance for enhancement contribution is increased to range from 5 to 20 nm from the MNP. This simplistic model does not take into account either quenching effects, or hot spot effects that can occur due to aggregation of MNPs especially at higher concentrations. This simple calculation also points to a possible future experiment in which MNP concentration would be increased by 2 orders of magnitude to achieve maximum potential enhancement, (from all the molecules present in the aerosol) which could result in overall fluorescence enhancement on the order of 100. By using spacer coated MNPs that increase the separation between the MNPs and the fluorophores the contribution from quenching could be reduced to increase the overall fluorescence enhancement.

Finally, it is not immediately clear why only quenching is observed in all emission bands for both the excitations for the 2 µm PSLs. One factor that could affect the differences observed between samples with coated and embedded MNPs is the lower strength of the absorption cross-section of gold MNPs when surrounded by air as compared to water/biological material [23]. Another factor could be the distribution of the MNPs on the aerosol. For surface coated aerosols, the number density is higher compared to their embedded counterparts that could cause higher agglomeration of MNPs as can be seen for PSL in Fig. 1 compared to tryptophan aerosol. The size of the aggregates will lead to different quenching and enhancement effects.

Fluorescence spectra of ovalbumin and tryptophan were also measured as a function of varying size and type of the MNPs. In ovalbumin samples, changing the size of gold MNPs from 10 nm to 50 nm in diameter resulted in similar behavior: enhancement was observed for all emission bands using 355 nm excitation, and when using 266 nm excitation, enhancement was observed for longer emission bands while quenching was observed for the two UV bands. However, for tryptophan samples, both 5 nm, and 50 nm sizes of gold MNPs were investigated. In the case of 5 nm MNPs, the presence of the MNPs resulted in no appreciable change in the fluorescence spectra, perhaps the diameter of the MNPs was too small to result in appreciable field enhancement. However, for 50 nm MNPs drastic quenching of the fluorescence was observed in all of the fluorescence bands for both excitations. Silver MNPs of 20 nm and 40 nm were also used with tryptophan, resulting in initial enhancement (10%) of the UV band for the 355 nm excitation and followed by quenching upon further increase of the concentrations of MNPs. For all the emission bands with 266 nm excitation, only quenching was observed.

6. Conclusions

We have demonstrated the ability to generate aerosols with the addition of MNPs, and to measure their effect on fluorescence emission by employing a real-time, single particle, on-the-fly system. To our knowledge this is the first report exploring in situ fluorescence enhancement/quenching by MNPs on aerosol particles. Aerosol particles composed of amino acids, proteins and PSL were generated both with and without the addition of small mass percentage quantities of 10 nm gold MNPs, either embedded throughout the aerosol particle, or on its surface in the case of the PSL particles. The fluorescence spectra of the individual aerosols were analyzed using both 266 nm and 355 nm excitations. Both enhancement and quenching effects were observed as a function of MNP concentration. The spectral decomposition was limited to discrete, broad bands and the MNP concentration dependence in the individual fluorescence bands, resulting in a shift in the shape of the fluorescence spectrum. On a fundamental level, the degree of observed enhancement or quenching is expected to be a function of several inter-dependent variables, including: the coupling efficiency of the excitation wavelength to MNP plasmon resonance, overlap of the particle material’s fluorescence emission spectra with the MNP plasmon band, spacing between MNPs and fluorophore molecules, number of MNP sites, and MNP agglomeration. The combined effects of these basic parameters determined our measured results. However we were unable to decouple these parameters in this initial study. By employing controls in future studies such as, using MNPs with specified spacer coatings, and choosing MNP materials with different plasmon resonance spectral positions relative to the excitation and emission bands, a better understanding of the dominant processes and their tradeoffs should be obtained. Ultimately, MNPs may be gainfully used to enhance the net fluorescence signal from individual aerosol particles for the purpose of hazardous material detection.

SEM images of the aerosols showed that MNP aggregation occurred in our samples that could result in either hot spot effects, or simply effectively act as larger MNPs with plasmon frequencies shifted to longer wavelength and increased scattering and absorption cross-sections. Methods to control aggregation via suspension composition or alternate droplet generation methods will be considered for future efforts. We also plan to explore this technique in terms of enhanced Raman spectra, especially with regard to the feasibility of acquiring compositional information from individual aerosol particles on-the-fly (in situ). If feasibility studies provide positive results, improved, real-time single particle measurements could be considered. Coating ambient, environmental aerosols in real-time with MNPs will be a challenge, one possible technique to achieve this might be by employing electrospraying technique [41]. Specific arrangements optimized for either Raman or fluorescence spectroscopy could lead to advanced aerosol sensors that identify individual particle compositions.