Study of Filtration in Polyethylene Fibers used in Masks for Protection Against SARS-CoV-2 via Luminescent Aerosolized Silicon Nanoparticles


 Commercial polyethylene (PE) fiber-based masks are currently used as personal filters for protection against various microorganisms. Due to the coronavirus (SARS-CoV-2) pandemic of 2020, the use of masks has become the critical mechanism in reducing the spread. The PE mask filter uses a sieve (geometry) in a spider web fashion to filter out microorganisms using Van der Waals atomic forces. However, the non-geometrical part of the filtration process is not fully understood. In this work, we utilized luminescent ultra-small silicon nanoparticles, which are Si-H or/and Si-OH terminated to examine how the filter operates at a chemical level. The particles were sprayed onto the fiber network by an atomizer and we used scanning electron microscopy (SEM), optical microscope and fluorescence spectroscopy under UV radiation. The images and measurements clearly showed that the Si nanoparticles bonded to the PE fiber network. The results were analyzed in terms of chemical bonding between Si nanoparticle and fiber. Our findings suggest that the PE fibers could act as a chemical filter via hydrogen or hydrolysis–based bonding or via Si-C bonding, which is complementary to their physical filtration ability via the geometric sieve process. Moreover, the results indicate that the filter would be effective against the novel coronavirus, not only by the geometrical sieve but can be additionally enhanced as a chemical filter by allowing the amine, carboxyl and hydroxyl function groups of the virus to attach to fibers via C-N and C-O bonds.


Introduction
In March 2020, the World Health Organization (WHO) declared COVID-19 a pandemic. The disease is caused by SARS-CoV-2 virus, which emerged in Wuhan, China in 2019 [1]. The virus is highly contagious and has a higher case fatality rate than other known respiratory viruses. The nature of transmission of the infection was not clear at the beginning and extreme precautions were in place when looking after patients with this infection. Similarly, the public and governments alike took different, often extreme, measures to halt the virus transmission. It is crucial to study the effectiveness of Personal Protective Equipment (PPE) including masks in preventing the transmission of the virus. [2][3][4][5][6]. Surgical masks are masks that are used in the healthcare setting to prevent the transmission of potentially pathogenic microorganisms from those who wear the masks to those who are at risk of contracting them by trapping the microorganisms that are shed into the respiratory secretions. Most masks are made by bonding nonwoven thermoplastic fabrics together, utilizing mechanical, heat or chemical methods.
Spunbond and melt-blown are the two most popular of those methods. Both techniques can produce variable fabrics such as polyamide, polyester, polyethylene, polypropylene, and polyurethane [7][8][9].
However, polyethylene (PE) is the most common as it is relatively cheap with low melt viscosity and easily processed [10]. Epidemiologic studies showed that respiratory secretions droplets from infected individuals that are generated through coughing and sneezing are the main method of viral transmission. [11][12][13][14][15][16][17]. It is important to understand the interplay of geometric and chemical characteristics of the surgical masks in filtering SARS-CoV-2 virus. [18][19][20][21]. To date, to the best of our knowledge, there are no studies that examined how PE filter masks capture SARS-CoV-2 virus. Despite the interest in the use of PE filter masks to isolate the virus, there has been no studies that discern the basic principle of operation of the filter, and the interplay between sieve-based and the chemical/biological-based operation. In this work, we utilize for the first time luminescent ultra-small silicon nanoparticles to examine the filtering process of the polyethylene (PE) mask filters. We examine some of microscopic implications of defects in commercial PE masking filters. Silicon nanoparticles have several advantages.
First, they are terminated by Si-H or/and Si-OH termination, which makes them chemically active through hydrogen bonding or hydrolysis-based bonding. Secondly, chemical attachment through covalent bonding is feasible since the hydrogen bond can be easily softened or cleaved via heating, irradiation, or oxidizers. Third, because they are ultrasmall (~ 3-nm) they are highly luminescent under UV/blue irradiation thus one can follow the path and landing through luminescence spectroscopy and microscopy. Fourth, because they are hydrophobic, clusters of diameters on the range of 100-300 nm across may form, which is similar to the virus size. Fifth, one may be able to use shifts in spectral distribution of the luminescence to discern the kind of bonding they make with the filter (physical vs. chemical, such as Si-C bonding). We used scanning electron microscopy (SEM) as well as optical microscopy to image the surface topography, bubble defects and the degree of roughness/randomness. The aggregation of silicon nanoparticles will be correlated with the defect distribution. The spectral distribution of the luminescence will be analyzed to discern the nature of attachment of the silicon nanoparticles to the fiber chemical vs. physical, namely of the defect cause changes in the molecular structures such as stripping of hydrogen and invitation of other active groups.

Si Nanoparticles
The Si nanoparticles in this study are produced by a chemical etching of boron doped 100 Si wafers [22].
We treated the wafer in a mixture of HF and hexachloroplatinc (H2PtCl6_H20) acid to deposit a platinum catalyst. The wafer is then etched briefly using HF/H2O2/methanol mixture. The wafer is then sonicated in a liquid of choice. Figure 1(a) gives the TEM image of a typical nanoparticle sample deposited on graphite grid; a histogram gives counts of 4, 56, 12, 9, and 2 at diameters of 1, 1.7, 2.2, 2.9, and 3.5 nm.
The first four produce wide emission bands at wavelengths of 410, 540, 570 and 600 nm respectively.
Unlike the sharp emission from nanoparticles of direct-bandgap material, the wide emission by single sized-particles here is due to the fact that the precursor material, i.e., silicon is an indirect bandgap material. This gives a total of 11 particles with the smallest size (blue-green emitters) and a total of 72 particles of larger particles (red-orange emitters) with a ratio of 11:72 or 85% purity of the larger particles. In other words, the chemical procedure produces a soup of particles with red luminescent particles (size of 2.9 nm in diameter) dominating the distribution. with an intensity ratio of the blue to the red being 0.6. The red-orange band covers the range of 540-720 nm, with band head at ~ 620 nm. The weaker blue green band covers the range ~ 400-480 nm corresponding to emission from the 1 and 1.67 nm diameter blue green emitting fraction of the particles seen in the TEM measurement. The quantum efficiency of emission of the red emitting particles is ~ 30 percent while that of the blue emitting particles is ~ 60 percent. The particles of sizes 1.7 and 1 nm emit radiation in the blue green part of the spectrum, while the 2.2, 2.9 and 3.5 nm emit in the red-orange range. The emission bands of all of the particles can be excited in the UV at wavelengths at ~ or below350 nm, but their emission can also be induced (shows local resonance structures) when excited at 3.4, 2.64, 2.4 and 2.1 eV respectively (Figure 1c). Those energies are the opening of their individual spatial quantum confinement bandgaps. Moreover, for excitation in the UV, the efficiency drops with the Stokes shift (energy difference between the excitation and emission photon energies). When examined the solution can be seen by the naked eye to be dominated by brightly red-orange luminescent under 365 nm irradiation as can be seen in the photoluminescent photo of the colloid given in Figure   1d. Figure 1d shows that when a soup of the particles solidifies on a silicon surface, same size particles assemble (crystalize) together, showing their distinct luminescence colors [23]  (c) emission spectra of 1, 1.67, 2.2, and 2.9 nm particles with excitation at their respective confinement bandgaps [23]. (d) The Si NPs solution can be seen to be dominated by brightly red-orange luminescence under 365 nm irradiation and the self-assembly on a surface of same size (emitting same band) nanoparticles

Chemical structure of the masks
The masks used are commercial products. The interaction of nanoparticles with microporous membranes or networks of polyethylene (PE) plastic fibers has been receiving concerted interest due to the ongoing SARS-CoV-2 challenge. The elementary structure of PE membranes is a repeating linear molecular structure or chain, the building unit of which is the hydrocarbon -CH2-CH2. The molecules can make solids in form of pellets, sheets, fibers, filaments. Those structures have found applications in a myriad of devices, an urgent one of which is a protective gear (masking) against virus entry into the lung system.
They are also widely used in laboratory applications for solid-liquid filtration/separation, medical devices, drug delivery devices, automotive industry, water filtration systems, and medical and surgical applications.
On a microscopic level, the importance of this material stems from the fact that the PE molecule is inert with strong molecular bonds and stable chemical resistance, as well as having a low drug or protein adsorption. In addition, it is nonpolar, hydrophilic and hydrophobic and self-sealing in water. It is a saturated (single bonds) hydrocarbon structure. As such, the chain macromolecules do not covalently link to each other but can associate with each other through weak atom-atom forces to make variety of material structures. Because of their inertness, they do not easily stick to surfaces, and as such they cannot be imprinted or bonded with adhesives without pretreatment. To achieve high strength joints, it is customary to use plastic welding. Moreover, their symmetric molecular structure in a solid form makes them show crystallinity and higher density albeit partial. As to the gas permeability in their solid form, polar gasses experience lower permeability than oxygen, carbon dioxide and flavorings, which can pass it easily. In their fiber form, they can be made in cloth configuration and used as gas filters. Filters based on inert material, such as PE operate on mechanical or geometrical principles (sieve effect), which is one most commonly applied in air filters. The principle of the effect is that the particle to be filtered is larger than the gap between the fibers and therefore the filter traps it without interfering with the chemistry of the particles. To prevent build-up, the filter must be cleaned regularly or discarded.
Replacement is only required when the filter can no longer release all the dirt that it holds.
The types of masked used in this work are highlighted in Table 1. They were compared to the lack of a mask barrier (no mask control)

Spray Deposition
We first studied the effect of aerosolized deposition of Si nano. The aerosolized droplets pattern from the spray bottle was first optimized by spraying a Si wafer under UV irradiation as shown in Figure 2. Using the spray deposition and the Si nano, the mask filter process will be studied. We used a spray NPs dispersion was added to 10 mL of IPA, sonicated for two hours to ensure a homogenous solution.
Before testing, spray bottle nozzles were adjusted to produce cloud and jet-propelled droplets that match the visual architecture of droplet formation described in [24,25].
Typical pump spray dispensers produce various liquid droplet dynamic phases occurring within a single spray [26], which reproduces results in a wide arrange of droplet sizes with a distribution range between 20 and 900 µm [27]. For comparison, the size of droplets in a typical sneeze range between 40 and 900 µm [28]. The spray bottle ejects fluid with pressures that can reach a pressure of 10 psi to create a short burst of microdroplets. In context, the pressure during a sneeze is 1 psi in the trachea and 2.6 psi in the mouth [29][30][31]. In this work, one spray pass ejects 250 µL of fluid per spray, therefore four consecutive spray passes, delivered at 1 pump/second, constitute an exposure of 1 ml of fluid in 4 s.
Since the nanoparticle colloid has a distribution of nanoparticle sizes, we expect the nanoparticles, if dried, to cluster with similar particles tending to self-assemble. Thus, the smaller blue-green particles may go together while the larger nanoparticles are expected to go together. When the particles in a colloid are sprayed onto a substrate using a jet/nozzle using differential pressure, they form droplets that encapsulate particles of similar size. In the present atomization, we expect approximately two components: red-orange emitting droplets and blue-green emitting droplets. The volatile liquid may evaporate in flight to a large degree, allowing the particles to land on the substrate.

Setup with Mask
To simulate the function of mask barriers, we placed selected masks framed between the spray bottle and 3×3 cm Si wafer piece mounted over a foam surface and above the table surface, through which droplets were sprayed as depicted in Figure 3(a,b,c). The dispenser nozzle tip was placed perpendicular to the masks at a horizontal distance of 10 cm mask which is around half the distance between the nostrils and vocal cords (~15-20 cm) [32], or the lip-to-carina distance ~20-25 cm [33] in humans.
Quantification represents droplets that pass through the masks and that land on Si pieces. All testing conditions were carried out at constant ambient conditions.

Results
The mask fibers were studied using several diagnostic procedures including optical, luminescent and electron spectroscopy. It was imaged using scanning electron microscopy (SEM), fluorescence and optical microscopy, and luminescent spectroscopy. Diagnostics were made while the particles were in the colloid and after landing on the fibers.

Electronic imaging
Since the fibers are non-conducting the resolution of SEM is limited. Standard use of deposition of a thin film of gold palladium on the fibers to enhance the resolution was not used to avoid quenching of the luminescence of the nanoparticles. The topographical characteristics of these samples were not

Luminescent imaging
Luminescence studies included imaging, microscopy, and spectroscopy.

Photoluminescence spectroscopy
The nanoparticles on the masks were examined using luminescence spectroscopy to determine the   As discussed above the spectrum shows a red-orange and a blue-green band. We make several observations. First, we see a relative drop in the red to blue ratio upon deposition on the mask.
After spraying, the red band intensity becomes smaller than that of the blue band. Second, upon deposition we see a drop in the red wing of the red band relative to its blue wing, which causes an apparent shift in the center of gravity of the band towards the blue by nearly 35 nm. Third upon deposition we see no change in the structure of the blue band hence no shift. On the average, the nanoparticle coating appears to have a pink tint compared to the red tint from the colloid indicating the increase of the blue band strength relative to the red band. The discontinuity in the spectra at is due to a filter change.
It is interesting to isolate the red luminescence band by exciting at 488 nm. At this wavelength, the photon energy is below the excitation energy of the blue band. Figure 8 gives the corresponding emission as a function of wavelength. FFP2 is a surgical mask made of three layers: a filter layer made of a melt-blown polymer, most commonly polypropylene placed between two non-woven fabric layers.
The melt-blown material is the material preventing the microbes from exiting the mask. The position of the measured red-orange band is in reasonable agreement with the band excited using UV radiation at given in Figure 7. The mask is a non-surgical mask, which is made from stretch nylon fabric. The spectrum is found to be blue shifted slightly relative to the polymer mask.

Figure 8
The PL spectra of coated FFP2 Mask and Cleanroom Mask at 488nm

Analysis and discussion
First, the above results show that there is a change in the red part of the emission spectrum of the nanoparticles when they are deposited on the mask. Such change in the spectral distribution may shed some information on the nature of attachment of the silicon nanoparticles to the fibers. For instance, the spectral change suggests that it is more probable that the deposition is of a chemical nature than of a physical nature. For the particles to chemically attach, there must be some active sites (defects) since the PE structure is inert. This may point to the presence of defect sites on which the nanoparticles may be attached. Defects may be created by striping some of the inert hydrogen termination off the fibers, making them active sites, which may also be followed with substitution by active groups. This effectively result in changes in the molecular structures of the fibers.
Defects may arise during manufacturing or due to effect of external agents. External agents can affect PE in general and fibers in particular. Weather of mechanical, chemical or radiation origin they are very important for their use in devices. First, the manufacturing process of commercial products is not fully controlled at the microscopic or molecular level. Bubbling during the extrusion process may introduce high mechanical strain or interruptions, which may introduce microscopic defects as was shown in Figure 4. Such defects may include those due to severing hydrogen and creation of dangling bonds, which makes the structure amenable to environmental attack. Although PE fibers in general are not attacked by strong acids or strong bases and are also resistant to gentle oxidants and reducing agents, interaction with strong peroxides can soften or severe hydrogen termination. Light can also create defects in PE. For instance, hydrogen termination may be softened or severed by exposure to strong light / UV.
In fact, PE is known to become brittle when exposed to sunlight due to cleavage of hydrogen and would crumble into small pieces. To alleviate this feature, carbon black is usually used to intercept UV and as such act as a UV stabilizer. The dangling bonds due to cleavage can also be easily saturated with oxygen or hydroxyl groups, resulting from environmental interactions, which may compromise the inert nature of the filter but allow chemical attachment to other agents.
We now discuss some plausible processes that may proceed at defect sites between the fibers as well as between external species and the fibers. An interaction of defect on a certain fiber can occur with another nearby fiber such that the two fibers are crosslinked through C-C bonding as shown in Figure 9. The present results have shown Si nanoparticles bond to the fibers of the various masks examined, acting as an effective barrier. We now discuss a plausible process for the bonding. Silicon nanoparticles may be partially oxidized resulting in termination by the OH function group (Si-OH) without compromising their luminescence. In this case, hydrogen bonding takes place with the fiber when the system is dried.
Also, the Si-OH of the nanoparticle can react with the C-H bond of the fiber to form the Si-C bond. In addition, as mentioned above, under UV irradiation the hydrogen bond is softened or cleaved in both the nanoparticles and the polyethylene fiber. A similar interaction allows the bare silicon and the bare carbon atoms in PE to form a Si-C bond as shown in Figure 10(a). The same nanoparticle could interact with another PE chain, essentially acting as a secondary cross linker as shown in Figure 10(b). The formation of the Si-C bond has been examined previously in 1-nm particles using first principle atomistic simulations and calculations. It was shown to increase the band gap of the particle because a C atom tends to repel the electrons to the inside of the particle. This may cause a shift in the center of the nanoparticle emission band towards the blue. The above spectral results given in Figure 7 show that the emission intensity ratio of the blue-green band and the red-orange band changes upon deposition on the mask. In the colloid, the intensity ratio of the blue to the red is 0.6. On the mask the ratio is 1.8. Since smaller nanoparticles, are responsible for the blue band, whereas larger nanoparticles (red emitting) are responsible for the red-orange band, one can conclude that the number ratio changes upon deposition. This may indicate that the droplet having smallest nanoparticle clusters fly longer in the air compared to those having the larger particles. It is also plausible that the change in the ratio is due to variations in the charge exchange processes between the chemically attached species, due to variations in the energy scales in the different sized particles. The confinement bandgaps of the blue and green emitting particles are 3.4 and 2.64 eV, compared to 2.4 and 2.1 eV for the orange and red emitting particles respectively. One effect that one can mention is that a chemical attachment of the red emitting particles may make them more susceptible to luminescence quenching and reduction of quantum efficiency.
The results may have implication on the filtering ability against SARS-CoV-2. Based on this, the ability of the fiber to filter out corona virus is interesting to address. The result shows the ability of the PE to chemically bond to Si-H or Si-OH. UV or light exposure can be used to photocatalyze binding of the nanoparticles to the fibers. Filtering SARS-CoV-2 presents specific active chemistry and challenges.
The protein molecules that constitute the spike protein that facilitate the entry of the virus into human cells is mostly amino acids having amine (NH2) and carboxyl (COOH) active functional groups. The above reported results using the silicon nanoparticles may suggest that the fiber can also be an effective SARS-CoV-2 filter not only by the geometrical sieve process but also as a chemical filter by allowing the amine, carboxyl and hydroxyl function groups of the virus to attach to fibers. Figure 11 shows a plausible mechanism for PE to attach chemically to the amine and the carboxyl groups in the virus via the formation of C-O and C-N covalent bonds. This is indicative of the virus being drawn. In addition, the protocols to disrupt the virus in human dwelling including the anti-virus action of oxidizers (H2O2), UV, or alcohol may also be useful in activating the filters against the virus. Those agents tend to break many of the weak linkages, or bonds (e.g., hydrogen bonds) (denaturation), within the protein molecules that constitute the casing or wall as well as the entry mechanism of the virus into cells. Such action results in a looser, more random structure; most of which are insoluble, rendering it ineffective (destroyed). Adding water to alcohol catalyzes the breakage of bonds hence the destruction of the virus. Hydrogen peroxide, a strong oxidizer produces hydroxyl free radicals that attack membrane lipids, DNA, and other essential cell components in a virus. Thus, those chemical agents are also expected to affect the filter by allowing PE chains to crosslink with each other using strong covalent bonding. As discussed above, peroxide, UV and electron irradiation can soften or break hydrogen bonds in PE such that it gets activated to cross-link chains of PE through the formation of C-C bonds. Moreover, it activates attachment of PE to the amine (NH2) or carboxyl (COOH) groups of the spike proteins of the SARS-CoV-2.

Conclusions
We utilized droplets encapsulating luminescent ultra-small silicon nanoparticles to examine how commercial polyethylene (PE) fiber-based masks work as personal filters for protection against various microorganisms. The particles have different sizes, wavelength (color) of emission, as well as Si-H or/and Si-OH termination. The nanoparticles were sprayed onto the fiber network by an atomizer placed at a certain distance and UV-induced fluorescence was used to observe their flight from the atomizer source to the mask target. The studies show that smaller nanoparticles are more efficient at reaching the target, and that the PE fibers act as a chemical filtering device via hydrogen or hydrolysis-based bonding or via Si-C bonding in addition to the geometrical sieve filtering. Moreover, the results indicate that the filter would be effective against the novel coronavirus, not only as a geometrical sieve filter but additionally as an enhanced chemical/bio filter via attachment of the amine, carboxyl and hydroxyl function groups of the virus to fibers via C-N and C-O bonds.