Fundamental protective mechanisms of face masks against droplet infections

Abstract

Many governments have instructed the population to wear simple mouth-and-nose covers or surgical face masks to protect themselves from droplet infection with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in public. However, the basic protection mechanisms and benefits of these masks remain controversial. Therefore, the aim of this work is to show from a fluid physics point of view under which circumstances these masks can protect against droplet infection. First of all, we show that the masks protect people in the surrounding area quite well, since the flow resistance of the face masks effectively prevents the spread of exhaled air, e.g. when breathing, speaking, singing, coughing and sneezing. Secondly, we provide visual evidence that typical household materials used by the population to make masks do not provide highly efficient protection against respirable particles and droplets with a diameter of 0.3–2 μm as they pass through the materials largely unfiltered. According to our tests, only vacuum cleaner bags with fine dust filters show a comparable or even better filtering effect than commercial particle filtering FFP2/N95/KN95 half masks. Thirdly, we show that even simple mouth-and-nose covers made of good filter material cannot reliably protect against droplet infection in contaminated ambient air, since most of the air flows through gaps at the edge of the masks. Only a close-fitting, particle-filtering respirator without an outlet valve offers good self-protection and protection against droplet infection. Nevertheless, wearing simple homemade or surgical face masks in public is highly recommended if no particle filtrating respiratory mask is available. Firstly, because they protect against habitual contact of the face with the hands and thus serve as self-protection against contact infection. Secondly, because the flow resistance of the masks ensures that the air stays close to the head when breathing, speaking, singing, coughing and sneezing, thus protecting other people if they have sufficient distance from each other. However, if the distance rules cannot be observed and the risk of inhalation-based infection becomes high because many people in the vicinity are infectious and the air exchange rate is small, improved filtration efficiency masks are needed, to take full advantage of the three fundamental protective mechanisms these masks provide.

Full text

Journal of Aerosol Science 148 (2020) 105617
Available online 28 June 2020
0021-8502/© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
Fundamental protective mechanisms of face masks against
droplet infections
Christian J. K€
ahler
*
, Rainer Hain
Institute of Fluid Mechanics and Aerodynamics, University of the Bundeswehr Munich, Werner-Heisenberg-Weg 39, 85577, Neubiberg, Germany
ABSTRACT
Many governments have instructed the population to wear simple mouth-and-nose covers or surgical face masks to protect themselves from droplet
infection with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in public. However, the basic protection mechanisms and benets
of these masks remain controversial. Therefore, the aim of this work is to show from a uid physics point of view under which circumstances these
masks can protect against droplet infection. First of all, we show that the masks protect people in the surrounding area quite well, since the ow
resistance of the face masks effectively prevents the spread of exhaled air, e.g. when breathing, speaking, singing, coughing and sneezing. Secondly,
we provide visual evidence that typical household materials used by the population to make masks do not provide highly efcient protection against
respirable particles and droplets with a diameter of 0.3–2
μ
m as they pass through the materials largely unltered. According to our tests, only
vacuum cleaner bags with ne dust lters show a comparable or even better ltering effect than commercial particle ltering FFP2/N95/KN95 half
masks. Thirdly, we show that even simple mouth-and-nose covers made of good lter material cannot reliably protect against droplet infection in
contaminated ambient air, since most of the air ows through gaps at the edge of the masks. Only a close-tting, particle-ltering respirator offers
good self-protection against droplet infection. Nevertheless, wearing simple homemade or surgical face masks in public is highly recommended if no
particle ltrating respiratory mask is available. Firstly, because they protect against habitual contact of the face with the hands and thus serve as
self-protection against contact infection. Secondly, because the ow resistance of the masks ensures that the air remains close to the head when
breathing, speaking, singing, coughing and sneezing, thus protecting other people if they have sufcient distance from each other. However, if the
distance rules cannot be observed and the risk of inhalation-based infection becomes high because many people in the vicinity are infectious and the
air exchange rate is small, improved ltration efciency masks are needed, to take full advantage of the three fundamental protective mechanisms
these masks provide.
1. Introduction
At present, humanity is threatened by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic. The risk of
severe infection with the virus depends heavily on physical factors of the infected persons and the quality of the medical system.
According to a recent study the estimated infection fatality ratio (IFR), averaged over all age-groups including those who don’t have
symptoms, is between 0.2% and 1.6% with an average of 0.66% (Verity et al., 2020). These numbers look small, and the fatality risk
may seem acceptable, and therefore the danger is often marginalized. This is surprising considering that the Apollo crew, the space
shuttle astronauts and the Allied soldiers during the 2003 Iraq war took a deadly risk of this magnitude. Only very few people take such
risks voluntarily and with full consciousness. For comparison, the lethal risk of a fatal accident with a commercial aircraft was
1:7700000 in 2008 and even such a small risk is not taken by some people. Considering that the IFR of the seasonal u is about
0.04–0.1% (Centers for Disease, 2010) or even much lower (Wong et al., 2013) the mortality rate of SARS-CoV-2 appears to be
signicantly higher than for inuenza u. Although the numbers for SARS-CoV-2 are quite preliminary and the estimates may drop
* Corresponding author.
E-mail address: christian.kaehler@unibw.de (C.J. K€
ahler).
Contents lists available at ScienceDirect
Journal of Aerosol Science
journal homepage: http://www.elsevier.com/locate/jaerosci
https://doi.org/10.1016/j.jaerosci.2020.105617
Received 20 May 2020; Received in revised form 15 June 2020; Accepted 16 June 2020

Journal of Aerosol Science 148 (2020) 105617
2
over time (Verity et al., 2020, Faust & del Rio, 2020) it is quite clear that the strategy of herd immunization of the population is not an
option, as the number of victims would be far too high. Great hopes for coping with the pandemic currently rest on the development of
a vaccine. Unfortunately, it is completely uncertain when an effective and well-tolerated vaccine will be generally available to contain
the pandemic. Drugs such as Chloroquine, Remdesivir, Lopinavir and Ritonavir are also considered to be great sources of hope in the
ght against the coronavirus disease 2019 (COVID-19) (Grein at al., 2020). However, even if one of the drugs should prove to be
effective, there is no guarantee that the drug can be made available to the world population in sufcient quantities. In addition, it is
possible that, despite the use of drugs, going through a severe course of disease can lead to lifelong neuropsychiatric sequelae (Troyer,
Kohn, & Hong, 2020; Zandifar & Badrfam, 2020) or cause other diseases (Ackermann et al., 2020; Varga et al., 2020).
Containing the pandemic is therefore the only viable way to quench the spread of the virus. But containing the pandemic is a
difcult task as about 44% of SARS-CoV-2 infections are caused by people with a presymptomatic and asymptomatic course of
infection (He et al., 2020). Therefore, due to the absence of symptoms, many people do not know that they are infected and are
spreading the virus and these people make it very difcult to trace the transmission chains. Furthermore, about 10% of infected people
are responsible for 80% of infections (Kupferschmidt, 2020; Lloyd-Smith et al., 2005). People who have many social contacts at work
or in their private lives and who do not protect themselves and others sufciently by observing the rules of distance and hygiene, or
who consider the risk of the virus to be low, appear to be a serious problem in the actual pandemic. For these reasons, the government
must act at various levels to avert great harm to the population. The effectiveness of the containment strategy depends on:
1. How societies are able to protect themselves personally against infection through hygiene, social distance and technical aids such as
protective masks, glasses, gloves.
2. How well the infrastructure is in place to identify the infection chains and effectively contain the spread, e.g. through mobile data
collection, isolation or a lockdown.
3. How well the seriously ill can be treated in hospitals.
In view of these prospects, it seems necessary for the time being to prevent the spread of the virus and to treat those infected as well
as possible. In order to ensure the latter, the capacities of the health system must not be overloaded. But it is clear that this condition
means that the pandemic will last for years without a vaccine
1
. To not overload the medical systems, governments are pursuing the
concept of containment by means of a lockdown because it proved successful in St. Louis during the Spanish Flu of 1918. This approach
is quite effective when the population obeys the rules, but the impact on the state, economy and society is devastating when the
lockdown lasts longer than a few weeks. Therefore, this concept is not a viable way to contain the pandemic in the long term.
Consequently, it is necessary to ght the infection where it occurs.
Understanding the transmission pathways is the key to nding effective measures to block the infection and to reliably protect
healthcare workers and the population. Contact infection were initially assumed to be the main transmission route of SARS-CoV-2.
Today, hygiene measures and the avoidance of shaking hands effectively prevent this path of infection. Droplet infection is
currently assumed to be the main transmission route over short distances (Wang et al., 2020). Since this path of infection is via the air,
the rules of distance are effective (Soper, 1919 and Wells et al., 1936). But it is also known that SARS-CoV-2 can remain infectious in
aerosols for more than 3 h, at least under laboratory conditions at high humidity (van Doremalen et al., 2020 and Pyankov et al., 2018).
It is therefore conceivable that infections can also occur under special conditions over long distances, provided that the local virus
concentration reaches the minimum infection dose due to poor air exchange in rooms.
A signicant proportion of the aerosol exhaled by humans has a diameter of less than 10
μ
m (Johnson et al., 2011) when breathing,
speaking, singing and coughing. It is also known that the size and number of droplets increases with the volume of the voice (Asadi
et al., 2019; Loudon & Roberts, 1968) and it is known that upper respiratory tract diseases increase the production of aerosol particles
(Lee et al., 2019). Water droplets of this size evaporate within a few seconds at normal humidity (Liu et al. 2019 and Rensink, 2004).
Droplets with a diameter of 10
μ
m for instance are evaporated after about 1 s at 50% relative humidity and larger droplets sink quickly
to the ground and evaporate (Marin et al., 2016, 2019; Rossi et al., 2019). If the viruses are released as “naked” viruses together with
the salt after the droplets have evaporated, the spatial concentration decreases rapidly over time, as the viruses no longer move in a
correlated manner but quickly separate due to the chaotic turbulent ow motion. The viral load thus decreases rapidly in time and
space, making infections over long distances or long periods of time increasingly unlikely. For this reason it is most important to
understand the transmission of the virus over short distances.
Hygiene regulations and social distancing are very effective in blocking short distance infections. During the lockdown, the distance
rules can usually be adhered to, but what happens when the actual lockdown is over and the people meet again in conned spaces?
Then additional effective and efcient protection is essential to stabilise infection rates. Since the viruses are spread by contact and
droplet infection, technical devices are required that effectively intervene in the chain of infection and effectively block infection. An
effective protection is the respiratory mask as known since 100 years (Soper, 1919). The SARS outbreak in Hong Kong suggested that
the use of simple face masks may have contributed to an overall reduction in the incidence of viral respiratory infections (Leung et al.,
2003; Lo et al., 2005). Another study has shown that even a simple surgical mask can effectively reduce SARS infection (Seto et al.,
2003). These results are supported by recent articles Leung et al., 2020b, Howard et al., 2020, Chu et al., 2020 and Zhang, Lib, Zhang,
Wang, & Molina, 2020).
1
It is clear that it is not enough to have a well tolerated and effective vaccine in large quantities. The vaccine must also be accepted and used by
the world population.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
3
It was surprizing that for months, WHO, the CDC and many public health professionals in Europe advised against wearing face
masks unless someone has COVID-19 or cares for someone who has COVID-19 (Feng et al., 2020; Leung et al., 2020a). This recom-
mendation was based on three allegations. First, it was said that there is no scientic evidence that face masks can protect against
droplet/aerosol infections. Second, it was argued that the population will not be able to wear the masks properly. Third, the statement
that people will feel safe when wearing masks and then become careless and take risks was frequently made. At the same time, these
experts have stressed that health professionals urgently need face masks to protect themselves effectively. This contradiction has
created uncertainty among the population and called into question the credibility of the experts. It is a fact that particle ltration masks
are recognized as legal occupational safety equipment and that the wearing of these masks in contaminated areas is required by labor
law. There is therefore no doubt that these masks, when used correctly, provide effective protection within the specication range. The
effectiveness of simple mouth-and-nose covers and surgical masks is less well accepted. The International Council of Nurses (ICN)
estimates that, on average, 7% of all conrmed cases of COVID-19 are among healthcare workers (ICN, 2020). This illustrates that
surgical masks may not provide the reliable protection against droplet infection, as anticipated. It is therefore very important to
distinguish clearly between the different mask types when talking about their protective function. Unfortunately, this was not done
sufciently by the virologists and politically responsible persons in the initial phase of the pandemic. Also the second argument is
questionable. Why should the people of Western societies not be able to protect themselves as many people in East Asian countries have
long been doing? Many people in East Asian countries have already recognized through numerous pandemics that proper masks work
effectively. It does not seem right to regard the Western population as unteachable or even incapable. The third argument is also false,
because the opposite is true according to scientic studies (Prather, Wang, & Schooley, 2020; Scott et al., 2007 and Ruedl et al., 2012).
If people protect themselves personally, they have dealt with the danger and therefore they benet from the protection of the safety
device and from the less risky behaviour due to insight. The reason why these facts were not appreciated by the experts is due to the
attempt to prevent competition for protective masks between medical personnel and the public.
In the meantime, the general perception of the protective effect of face masks has become generally accepted. In the USA, the CDC
has changed its guidelines and recommended that the public wear fabric face masks. In other countries, too, it is now recommended to
protect themselves with suitable masks. However, it is recommended by governments and professionals to wear only simple mouse-
and-nose covers that can be manufactured by the people themselves or surgical masks to avoid distribution battles with medical staff
for certied and comfortable particle ltration masks. But the big question is, how effectively these homemade mouth-and-nose covers
and surgical masks can protect against droplet infection. The answer is highly relevant to guide public behaviour (Leung et al., 2020).
One study suggests that a surgical face mask and masks made of dense cotton fabrics apparently cannot effectively prevent the spread
of SARS-CoV-2 into the environment through the coughing of patients with COVID-19 (Bae et al., 2020; Klompas, Morris, Sinclair,
Pearson, & Shenoy, 2020). Another study suggests that any mask, no matter how efciently it lters or how well it is sealed, has
minimal effect unless used in conjunction with other preventive measures such as isolation of infected cases, immunization, good
respiratory etiquette and regular hand hygiene (Kwok et al., 2015). These ndings contradict the results in (Leung et al., 2003; Lo et al.,
2005; Seto et al., 2003). Due to the contradiction, it is understandable that experts in the media have expressed the opinion that there is
no scientic evidence for the effectiveness of masks and therefore the wearing of masks in public was not recommended for a long time.
The fallacy of politicians and virologists, however, was to generalize the results obtained with simple mouth-and-nose covers to all
masks without differentiation.
In order to clarify whether or to what extent these masks offer effective protection against droplet infection and to understand why
research results differ on this simple scientic question, we have carried out these tests. First, we analyse the ow blockage caused by
surgical masks when coughing, as this is essential for the protection of others and because coughing is a typical symptom of COVID-19.
Second, we qualify the effectiveness of different lter materials and masks to determine the protection ability against droplets. Finally,
we prove the effect of gap ows at the edge of surgical and particle ltrating respiratory masks. In contrast to the medical studies cited,
we apply engineering research methods of uid mechanics. The use of this research approach has several reasons: Firstly, the
detachment of droplets in the lungs and throat and their convective transport through the mouth into the atmosphere until inhalation
as well as the deposit and evaporation of droplets is a purely uid mechanical process. Secondly, the effective blocking of the ow with
suitable masks is a research subject of uid mechanics. Thirdly, the ltering of particles from an air stream with the aid of suitable
materials is also a purely uid mechanical problem as well as the gap ow. Finally, this approach also has the advantage that the results
are reproducible in a statistical sense, since the boundary conditions are well dened. We are not studying whether an infection really
occurs in a special case, but whether an infection is physically possible in general.
2. Materials and methods
In the rst sets of the experiments, outlined in section 3.1, the ow eld generated by coughing without and with a surgical mask is
examined as coughing sets the air strongly in motion and because coughing is a typical symptom of COVID-19. To measure the ow
eld quantitatively in space and time we use Particle Image Velocimetry (PIV) (Raffel et al., 2018). For the measurements a 8 m long
testing room with a cross section of 2 m �2 m was seeded with DEHS (Di-Ethyl-Hexyl-Sebacate) tracer particles with a mean diameter
of 1
μ
m (K€
ahler et al., 2002). DEHS was used as these droplets exist for several hours until they have evaporated. The tracer particles
provided by a seeding generator (PIVTEC GmbH, Germany) were illuminated in a light-sheet generated with a frequency doubled Nd:
YAG laser (SpitLight PIV 1000–15, InnoLas Laser GmbH, Germany). The light-sheet was oriented normal to the mouth opening and
parallel to the symmetry axis of the body and the longitudinal axis of the room. The light scattered by the tracer particles were recorded
with back illuminated scientic CMOS cameras (pco.edge 5.5, PCO AG, Germany) equipped with Zeiss Distagon T* lens with a focal
length of 35 mm and 50 mm. The triggering of the system components was achieved with a programmable timing unit (PTU X,
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
4
LaVision GmbH, Germany). The recorded series of images were evaluated with a commercial computer program (DaVis, LaVision
GmbH, Germany). These quantitative PIV measurements allow to determine the area that can be contaminated due to the exhaled air,
the velocity of the exhaled droplets and the turbulence properties of the ow.
In the second set of experiments, discussed in section 3.2, common household materials currently used by the population and some
medical staff to make simple masks at home were tested but also a surgical mask and a FFP3 mask to visualize their ltering properties.
The tested materials are given in Table 1. For the investigation, a test set-up was installed which largely fullled the ofcially pre-
scribed test conditions in Europe (DIN EN 149, 2009). The materials were installed one after the other in a xed position in front of the
inlet of a rectangular ow channel with a cross-section of 0.1 m �0.1 m, as shown in Fig. 1. The material was held in place with a
special clamping device that seals tightly to the duct to avoid leakage ows. To explore the ltering performance of the different
materials the movement of small aerosol droplets passing through the media was observed visually in front of and behind the lter
material with a digital camera. We only use droplets whose diameter is less than 2
μ
m, since the removal of the smallest droplets in an
air stream is the greatest challenge in mask development. If these droplets can be effectively ltered out effectively, then all droplets
larger than 2
μ
m can also be ltered. The droplets were generated from DEHS with an aerosol generator (AGF 2.0, Palas GmbH,
Germany). DEHS was used again as these droplets are long lasting. Consequently, bias errors due to evaporation effects can be
neglected. A Nd:YAG double-pulse laser (Evergreen 200, Quantel, France) was used to illuminate the droplets. The output beam was
fanned out with a few lenses to form a 1 mm thin light-sheet. The light-sheet was located in the middle of the ow channel parallel to
the ow direction as indicated in Fig. 1.
The scattered light emitted by the illuminated aerosol in the light-sheet plane was recorded with a highly sensitive PCO edge 5.5
sCMOS camera equipped with a Zeiss Distagon T* lens with a focal length of 50 mm. The triggering of the system components and the
data recording was realized again with the software DaVis from LaVision. The ow velocity was driven by the pressure difference
between the atmosphere and the ow box. The ow rate through the lter material was adjusted approximately according to the DIN
EN 149, 2009 test standard (90 L/min). The volume ow rate and the movement of the droplets through the lter material was
measured optically with high spatial and temporal resolution using PIV. To calculate the volume ow rate the average ow velocity
within the light-sheet plane in the ow channel was measured and it was assumed that this velocity is homogeneous over the cross
section of the channel. This assumption is justied as the ltering materials are homogeneous and the inow condition is constant
across the ltering material. With the know size of the cross section the volume ow rate can be calculated. The pressure drops
provided in Table 1 are calculated from the measured pressure drops and the volume ow rate. It is assumed that the pressure loss is
proportional to the square of the volume ow rate. The pressure drop across the lter material was measured with a pressure
transducer with an uncertainty of about 3 Pa (TESTO 480, Testo SE & Co. KG, Germany).
For the third set of experiments, analysed in section 3.3, simple ow visualizations using smoke were performed in order to
demonstrate the effect of the gap around the mask edge. A person exhaled air seeded with tracer particles while wearing surgical and
FFP2 masks.
3. Results
3.1. The effect of ow resistance (protection of others)
In the rst series of experiments, one person performed a single severe cough while the PIV system was measuring the ow eld
data. The video in the supplementary material shows the temporal evolution of the process. The results displayed in this subsection
show instantaneous velocity elds of various independent time-resolved ow eld measurements. Color-coded is the magnitude of the
local ow velocity and the vectors indicate the direction of the ow movement at a given time step. In areas where the ow movement
remains close to zero over the whole recording time (blue color), no droplets can penetrate as only the ow can move the particles to
other areas. Large droplets with a diameter of 1 mm or more, such as those produced when sneezing (Lok, 2016), can y ballistic over
long distances, and occasionally ballistic ying droplets are produced when certain sounds are spoken. But sneezing is not a typical
COVID-10 symptom so that this will not we considered here. The small droplets that are normally produced when breathing, speaking,
singing and coughing are immediately slowed down and then move with the ow velocity of the ambient air. It is therefore important
to study the air set in motion by exhalation. Furthermore, the small droplets are particularly dangerous because they can be inhaled
Table 1
Tested lter materials.
Material Surgical face mask Hygienic mask Toilet paper Paper towel Coffee lter
Pressure drop at 60 l/min [mbar] 0.2
a
0.3 0.3 1.8
Pressure drop at 90 l/min [mbar] 0.4
a
0.6 0.7 4.0
Material Microbre cloth Fleece Vacuum cleaner bag FFP3 mask with valve Halyard H600
Pressure drop at 60 l/min [mbar] 7.7 0.1 2.0
b
1.4
Pressure drop at 90 l/min [mbar] 17.2 0.2 4.5
b
3.1
a
Not measured.
b
Not available. Due to the valve and the resulting inhomogeneous ow eld, the volume ow rate could not be determined with the method
applied.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
5
deep into the lungs.
Fig. 2a shows that the spread of the exhaled air forms a cone like shape similar to a free turbulent jet (see video). The ow velocity is
reaching values up to 1 m/s near the mouth, but due to the widening of the cone caused by the turbulent mixing and entrainment
(Reuther & K€
ahler, 2020) the ow velocity decreases in streamwise direction. The widening of the area in motion reduces the viral load
signicantly with distance. A single strong cough sets the air in motion over a distance of less than 1.5 m in the experiments. Distances
of more than 1.5 m can be considered safe according to these results, since no droplets can reach such large distances when accelerated
by a single cough. However, if the cough lasts longer, greater distances can be achieved, as shown in Fig. 2b. For this reason, it is
important to dynamically increase the distance to a person if the coughing stimulus is about to last longer.
The results in Fig. 2c illustrate how the spread of the airow from the mouth during coughing is very effectively inhibited by a
surgical mask. Physically the mask ensures that the directional jet like air movement with high exit velocity from the mouth is con-
verted into an undirected air movement with low velocity behind the mask. This is because the exhaled air increases the pressure inside
the mask compared to the atmosphere outside, and the pressure difference creates a ow movement in all directions. This effect is of
utmost importance for limiting the virus load in the environment. The results show that even a simple mouth-and-nose cover or a
surgical mask can effectively protect other people in the vicinity because the mask prevents the droplets from spreading over a wide
area. A simple mask with sufcient ow resistance therefore provides very effective protection for people in the surrounding area when
infected and wearing the mask. Wearing a mask is therefore absolutely useful to protect others according to our quantitative
measurements.
Fig. 2d shows the spread of exhaled air when speaking. It can be clearly seen that a greater spread of the exhaled air appears than
when coughing with a mask. Consequently, wearing a mask during normal face-to-face conversations and of course also when talking
on the smartphone in a human environment is extremely useful to stop the transmission of the SARS-CoV-2 infection via droplets. It
must also be taken into account that persons with a presymptomatic or asymptomatic course of infection will infect other persons most
likely during face-to-face conversations. A mask will therefore make an effective contribution to suppressing this signicant path of
infection.
3.2. The effect of aerosol ltration (self protection)
In this section we want to nd out if the material of simple mouth-and-nose covers, surgical masks and FFP3 masks can protect the
user from droplet infection, if the surrounding air is contaminated with SARS-CoV-2. In this case, the mask material must have good
ltering properties to stop small droplets that typically occur when speaking, singing and coughing. Since large droplets are easily
ltered out by simple materials, we focus on small droplets in the range between 0.3 and 2
μ
m because they are produced in large
fractions when speaking, singing and coughing and they can penetrate deep into the lungs. The droplets were distributed approx. 400
mm in front of the lter materials. In order to make the motion of the droplets and the ltering ability of the materials clearly visible an
inhomogeneous droplet distribution was generated. The ow direction is from left to right and the ow state of the incoming air is
laminar. If the intensity of the scattered light emanating from the droplets is large in front of the lter material (left image) and close to
zero behind the lter material (right image), the droplets are almost completely ltered out through the material. If, on the other hand,
no signicant reduction in intensity can be detected behind the lter material, the lter effect is negligible. The area of the lter mount
and the channel edges are not shown in the following images, since no relevant ow and droplet information is visible in these areas.
The results presented are qualitative, but intended to be this way to provide readers with visual evidence of the particle penetration
Fig. 1. Schematic representation of the experimental setup.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
6
through different candidate lter media. A better impression of the lter efciency is obtained by viewing the second video in the
supplementary material.
The comparison of the two pictures in Fig. 3 (left) shows that almost all droplets pass the tested surgical face mask unhindered.
Consequently, this mask does not provide serious self protection against droplet infection. Only a mixing of the droplet distribution
takes place due to the porosity of the lter material. It is fatal that medical personnel are often so poorly protected by these masks. But
it is also fatal for patients if clinical staff with a presymptomatic or asymptomatic course of infection uses these masks.
Even worse than the surgical face masks is the hygiene mask, see Fig. 3 (right). This mask is designed for catching larger objects
such as hair and spook, but tiny droplets, such as those produced when talking, singing and coughing, cannot be ltered out of the air
stream by the hygiene mask. It should also be noted that the ow resistance of the hygiene mask is so low that even the protective
mechanism described in subsection 3.1 does not function effectively. Fig. 4 reveals the effectiveness of particle ltering with toilet
paper with 4 layers, paper towel, coffee lters, and microbre cloth which also offer no serious protection against droplets in this size
Fig. 2. Instantaneous ow eld when coughing over one breath without mask (a, top). Flow eld when coughing over a longer periods of time
without mask (b, middle). Flow eld when coughing over one breath with mask (c, lower left) and instantaneous ow eld when talking without
mask (d, lower right). The rst video in the supplementary material shows the animated sequences.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
7
Fig. 3. Effectiveness of particle ltering with the lter material of a surgical face masks (left) and a hygienic mask (right). The arrow indicates the
ow direction and lter position. The second video in the supplementary material shows the animated sequences.
Fig. 4. Effectiveness of particle ltration. Toilet paper (upper left), paper towel (upper right), coffee lter (lower left), microbre cloth (lower
right). The second video in the supplementary material shows the animated sequences.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
8
range. Only very large droplets are retained by these materials and therefore these materials are suitable for their intended use, but not
as lter material for small droplets. It is therefore strongly discouraged to make masks from these materials with the aim of protecting
oneself from infection.
Furthermore, a very strong eece was tested, which serves as a protective coating on ironing boards. The material is 4 mm thick,
completely opaque and has a pressure drop of about 35 Pa. However, a lter effect is not visible, as indicated in Fig. 5 (left). The droplet
clouds ow almost unltered through the eece. Even several layers of a dense fabric do not have a proper ltering effect on the
considered droplet sizes, which escape mainly when breathing, speaking, singing and coughing.
Good results could only be achieved with the material of a vacuum cleaner bag with ne dust lter properties, see Fig. 5 (right).
Despite the small droplets used in these tests, almost all droplets are reliably ltered out. Consequently, also no larger droplets will be
able to pass through the material. According to the manufacturer SWIRL, the material lters 99.9% of ne dust down to 0.3
μ
m
diameter. This vacuum cleaner bag with ne dust lter therefore has better ltering properties than all tested materials and masks and
even an FFP2 protective mask has poorer ltering properties, as it only has to lter out 94% of the ne dust down to 0.6
μ
m to meet the
specications (Uvex, 2020). The material of vacuum cleaner bags with ne dust protection is therefore very well suited as a
self-protecting mask if only the lter effect is considered. However, because vacuum cleaner bags are not certied clinical products,
they may contain unhealthy ingredients that kill bacteria and harmful bers that may leak from the bag material. It is therefore
uncertain whether this material is suitable in practice as a material for a respirator mask.
Fig. 6 (left) illustrates the ltering capabilities of an FFP3 mask under the test conditions. Nearly all droplets are ltered out as
expected. Therefore, this mask type is very well suited to protect people from an infection by means of aerosols even when the
environment is strongly contaminated with infectious droplets. Recently, some hospitals in the USA make use of Halyard H600 ma-
terial to protect their employers from aerosol infection. The test result of the material is displayed in Fig. 6 (right). It is clearly visible
that the ltering capacity of the material is not sufcient to protect people from infection by aerosols if the environment is contam-
inated with the SARS-CoV-2. Fig. 7 shows with different resolution microscopic images of the Halyard H600 material. It is composed to
bers but the density might not be sufcient to lter the particles used in our investigation. There are also tiny holes in the pockets
visible, which could be the reason why the aerosol passes through the material, as the ow resistance at the holes is low compared to
the other parts of the material.
The ow tests clearly show that apart from the vacuum cleaner bag and the FFP3 mask, the lter effect of the tested materials is not
sufcient to protect against droplet infection reliably if the environment is contaminated with SARS-CoV-2. Even masks routinely used
by medical staff in hospitals and doctor’s ofces have almost no signicant ltering effect on the droplet sizes typically produced when
breathing, speaking, singing and coughing. The results are therefore in good agreement with the results from (Leung et al., 2020 and
Davies et al., 2013 and Kwok et al., 2015. But why has wearing these masks been shown to provide effective protection against
infection with the virus in the SARS epidemic, as shown in (Leung et al., 2003, Lo et al., 2005 and Seto et al., 2003)? Because a mask is
important not only because of its ltering ability, but to limit the droplet propagation as discussed in section 3.1. So in combination
with distances these mask can protect if only a few people are infected in the surrounding. The results in (Bae et al., 2020; Leung et al.,
2020) are correct, but they do not consider the full performance of masks, but only a partial aspect. Therefore, the conclusions in the
articles are not universal. The ndings in (Leung et al., 2003; Lo et al., 2005; Seto et al., 2003) are understandable when the full
performance of masks in blocking infections is considered.
Unfortunately, wearing a simple mouth-and-nose cover may be less comfortable than wearing a particle ltering face mask. In
effect, this can promote a smear infection. Since all these transmissions of infection are possible in daily life, wearing a comfortable
mask is essential to block human-to-human transmission by smear and droplet infection. To ensure the best possible protection, a
particle ltering mask should be used if the number of infected persons in the environment and the viral load in the room is unknown.
At present, social distancing practices and universal masking seem to be the best methods of containing viral pandemic without stricter
lockdown policies and without vaccines.
Some recent studies show that even the simple materials we have tested have some ltering ability (Davies et al., 2013; Drewnick,
2020; Konda et al., 2020 and van der Sande et al., 2008), We do not question these results, although the pressure drops in one study is
anomalously low (see supporting information in Konda et al., 2020), but we state explicitly that a material that does not have an
adequate ltering ability equivalent to an FFP2/N95/KN95 mask cannot be recommended as a lter material for self-protection
Fig. 5. Effectiveness of particle ltration of a eece (left) and vacuum cleaner bag with ne dust lter (right). The second video in the supple-
mentary material shows the animated sequences.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
9
Fig. 6. Effectiveness of a FFP3 mask (left) and Halyard H600 material (right). The second video in the supplementary material shows the
animated sequences.
Fig. 7. Microscopic images of the Halyard H600 material with different resolution.
Fig. 8. Exhalation of air during light exhalation without physical exertion (left), heavy breathing comparable to physical exertion (middle) and
when coughing (right). Top row: Surgical mask. Lower row: Very simple FFP2 mask without exhalation valve. The rst video in the supplementary
material shows the animated sequences.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
10
against droplet infection. Statistically speaking, every loss of performance leads to an increase in the number of infected people and
thus to an increase in the number of death. It is therefore very dangerous to recommend materials with some ltering properties as
possible materials for self-protection masks. But there is another important aspect that will be discussed next.
3.3. The effect of gap ows on the mask performance
According to the previous section one might argue that a mouth-and-nose cover or surgical mask made of a good lter material
would provide good protection against infection when infected people are in the vicinity or the room is contaminated with viruses. But
that will not usually be the case. Air takes the path of least resistance. As these masks do not seal tightly enough with the face, droplets
can ow unhindered past the edge of the mask when inhaled and exhaled and reach the lungs or the environment. If the mask does not
t properly, this will even be the rule. This is illustrated in Fig. 8 were a person is exhaling air during an easy exhalation without
physical exertion (left), strong breathing during physical exertion (middle) and when coughing (right). The rst video in the sup-
plementary material shows the animated sequences.
4. Summary and conclusion
The analysis shows that it is very important do differentiate between mouth-and-nose cover, surgical mask and particle ltering
respirator mask because they differ substantial in their fundamental protection properties. Face masks can offer three fundamental
different kind of protection:
1. They effectively prevents a smear infection, as the wearers of the masks no longer perform their habitual grip on the face and thus
no longer bring the virus from the hand into the mouth or nose (Howard et al., 2020).
2. The ow resistance of the mask greatly limits the spread of viruses in the room. This signicantly reduces the risk of infection in the
vicinity of an infected person (protection of third parties).
3. The inhalation of droplets containing viruses can be prevented by using a tight-tting mask with particle ltering properties (self-
protection).
The rst fundamental protection mechanism can be reached by all face masks if they t well and sit comfortably. If not, the user will
touch the face even more than usual to correct the t of the mask. As this can increase the risk of smear infection, a good t of the mask
is very important. The rst and second fundamental protection mechanisms are fullled by all masks that have sufcient ow
resistance. If the mask is worn and a candle can easily be blown out despite the mask, the mask does not full this function and should
not be used. All three fundamental protection mechanisms can be only achieved with FFP2/N95/KN95 or better particle ltering
respirator mask.
Typical materials currently used by the public to build masks reduce the risk of smear infection and effectively prevent the
widespread spread of viruses in the environment. Therefore, the use of these mouse-and-nose covers and surgical masks are very
important to prevent smear infection and droplet infection to others if the distance is not too close. As these masks do not have a
signicant particle-ltering protective effect against droplets that are typically produced when breathing, speaking, singing, coughing
and sneezing they should not be used if the environment is contaminated, like in hospitals, even when the distance rules are followed.
To achieve effective self-protection in a virus-contaminated environment, masks with particle ltering properties (FFP2/N95/KN95)
are absolutely necessary from our point of view. If a large number of infected persons are present and distance rules cannot be ach-
ieved, a very good particle ltration mask (FFP3 or better) is strongly recommended.
If these general rules are followed and all people use suitable particle-ltering respirators correctly, the transmission of viruses via
droplets/aerosols can be effectively prevented. Otherwise, these types of masks would never have received certication, nor would
they be a core component of the personal protective equipment in hospitals and other environments. Therefore, proper face masks can
save lives while maintaining social life and securing the economy and the state.
But universal masking alone is not enough for two reasons: First, many people are not very good at following rules consistently.
Therefore, it is advisable to observe the rules of hygiene and distance and to be careful even when wearing a mask. In the event of a car
accident, the occupants are also protected by various devices (bumpers, crumple zone, safety belts, airbags, head and legroom,
autonomous assistance systems, …). Second, some people are extremely bad at following rules, either because they do not want to or
because they simply cannot. These people can become super spreaders. Therefore, the early detection of sources of infection and their
isolation remains important beside universal masking and the rules of hygiene and distance.
Acknowledgement
The authors would like to thank Stefan Ostmann for conducting the mask experiments presented in Section 3.3 and Amirabas
Bakhtiari for taking the microscopic images in Fig. 7.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.org/10.1016/j.jaerosci.2020.105617.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
11
References
Ackermann, M., Verleden, S. E., Kuehnel, M., Haverich, A., Welte, T., Laenger, F., Vanstapel, A., Werlein, C., Stark, H., Tzankov, A., Li, W. W., Li, V. W., et al. (2020).
Pulmonary vascular endothelialitis, thrombosis, and angiogenesis in Covid-19. New England Journal of Medicine. https://doi.org/10.1056/NEJMoa2015432. May
21.
Asadi, S., Wexler, A. S., Cappa, C. D., Barreda, S., Bouvier, N. M., & Ristenpart, W. D. (2019). Aerosol emission and super emission during human speech increase with
voice loudness. Nature Scientiftic Reports, 9, 2348. https://www.nature.com/articles/s41598-019-38808-z.pdf.
Bae, S., Kim, M.-C., Kim, J. Y., Cha, H.-H., Lim, J. S., Jung, J., Kim, M.-J., Oh, D. K., Lee, M.-K., Choi, S.-H., Sung, M., Hong, S.-B., Chung, J.-W., & Kim, S.-H. (2020).
Effectiveness of surgical and cotton masks in blocking SARS-CoV-2: A controlled comparison in 4 patients. Annals of Internal Medicine. https://doi.org/10.7326/
M20-1342. published online on April 6.
Centers for Disease C, Prevention. (2010). Estimates of deaths associated with seasonal inuenza ––– United States, 1976–2007. MMWR Morb Mortal Wkly Rep, 59
(33), 1057–1062.
Chu, D. K., Akl, E. A., Duda, S., Solo, K., Yaacoub, S., & Schünemann, H. J. (2020). Physical distancing, face masks, and eye protection to prevent person-to-person
transmission of SARS-CoV-2 and COVID-19: A systematic review and meta-analysis. The Lancet, 395, 1973–1987. https://doi.org/10.1016/S0140-6736(20)
31142-9. published online on June 1.
Davies, A., Thompson, K.-A., Giri, K., Kafatos, G., Walker, J., & Bennett, A. (2013). Testing the efcacy of homemade masks: Would they protect in an inuenza
pandemic? Disaster Medicine and Public Health Preparedness, 7, 413–418. https://doi.org/10.1017/dmp.2013.43.
DIN EN 149. (2009). Respiratory protective devices – ltering half masks to protect against particles – requirements, testing, marking. German version EN, 149, 2001þ
A1, Beuth Verlag GmbH, 10772 Berlin.
Doremalen van, N., Bushmaker, T., Morris, D. H., Holbrook, M. G., Gamble, A., Williamson, B. N., Tamin, A., Harcourt, J. L., Thornburg, N. J., Gerber, S. I., Lloyd-
Smith, J. O., de Wit, E., & Munster, V. J. (2020). Aerosol and surface stability of SARS-CoV-2 as compared with SARS-CoV-1. New England Journal of Medicine, 382,
1564–1567. https://doi.org/10.1056/NEJMc2004973.
Drewnick, F. (2020). Abscheideefzienz von Mund-Nasen-Schutz Masken, selbstgen€
ahten Gesichtsmasken und potentiellen Maskenmaterialien. https://www.mpic.
de/4646696/ltermasken_zusammenfassung_08_04_2020_v3_nal.pdf. called on May 28.
Faust, J. S., & del Rio, C. (2020). Assessment of deaths from COVID-19 and from seasonal inuenza. JAMA Intern Med. https://doi.org/10.1001/
jamainternmed.2020.2306. Published online May 14.
Feng, S., Shen, C., Xia, N., Song, W., Fan, M., & Cowling, B. J. (2020). Rational use of face masks in the COVID-19 pandemic. The Lancet Respiratory Medicine, 8,
434–436. https://doi.org/10.1016/S2213-2600(20)30134-X.
Grein, J., Ohmagari, N., Shin, D., Diaz, G., Asperges, E., Castagna, A., & Oda, R. (2020). Compassionate use of Remdesivir for patients with severe COVID-19. New
England Journal of Medicine, 382, 2327–2336. https://doi.org/10.1056/NEJMoa2007016. Published on April 10.
He, X., Lau, E. H. Y., Wu, P., Deng, X., Wang, J., Hao, X., Chung Lau, Y., Wong, J. Y., Guan, Y., Tan, X., Mo, X., Chen, Y., Liao, B., Chen, W., Hu, F., Zhang, Q.,
Zhong, M., Wu, Y., Zhao, L., … Leung, G. M. (2020). Temporal dynamics in viral shedding and transmissibility of COVID-19. Nature Medicine, 26, 672–675.
https://doi.org/10.1038/s41591-020-0869-5.
Howard, J., Huang, A., Li, Z., Tufekci, Z., Zdimal, V., van der Westhuizen, H., von Delft, A., Price, A., Fridman, L., Tang, L., Tang, V., Watson, G. L., Bax, C. E.,
Shaikh, R., Questier, F., Hernandez, D., Chu, L. F., Ramirez, C. M., & Rimoin, A. W. (2020). Face masks against COVID-19: An evidence review. Preprints. not peer
reviewed, Published on April 12 www.preprints.org https://doi.org/10.20944/preprints202004.0203.v1.
International Council of Nurses (ICN). (2020). https://www.icn.ch/news/more-600-nurses-die-covid-19-worldwide. Called on June 10.
Johnson, G. R., Morawska, L., Ristovski, Z. D., Hargreaves, M., Mengersen, K., Chao, C., Wan, M.-P., Li, Y., Xie, X., Katoshevski, D., & Corbett, S. (2011). Modality of
human exposed aerosol size distributions. Journal of Aerosol Science, 42, 839–851. https://doi.org/10.1016/j.jaerosci.2011.07.009.
K€
ahler, C. J., Sammler, B., & Kompenhans, J. (2002). Generation and control of tracer particles for optical ow investigations in air. Experiments in Fluids, 33, 736–742.
Klompas, M., Morris, C. A., Sinclair, J., Pearson, M., & Shenoy, E. S. (2020). Universal masking in hospitals in the Covid-19 era. New England Journal of Medicine, 382,
e63. https://doi.org/10.1056/NEJMp2006372.
Konda, A., Prakash, A., Moss, G. A., Schmoldt, M., Grant, G. D., & Guha, S. (2020). Aerosol ltration efciency of common fabrics used in respiratory cloth masks. ACS
Nano, 14, 6339–6347. https://doi.org/10.1021/acsnano.0c03252.
Kupferschmidt, K. (2020). Why do some COVID-19 patients infect many others, whereas most don’t spread the virus at all? Science. https://doi.org/10.1126/science.
abc8931. May 19.
Kwok, Y. L. A., Gralton, J., & McLaws, M.-L. (2015). Face touching: A frequent habit that has implications for hand hygiene. American Journal of Infection Control, 43,
112–114.
Lee, J., Yoo, D., Ryu, S., Ham, S., Lee, K., Yeo, M., Min, K., & Yoon, C. (2019). Quantity, size distribution, and characteristics of cough generated aerosol produced by
patients with an upper respiratory tract infection. Aerosol and Air Quality Research, 19, 840–853.
Leung, N. H. L., Chu, D. K. W., Shiu, E. Y. C., Chan, K.-H., McDevitt, J. J., Hau, B. J. P., Yen, H.-L., Li, Y., Ip, D. K. M., Peiris, J. S. M., Seto, W.-H., Leung, G. M.,
Milton, D. K., & Cowling, B. J. (2020b). Respiratory virus shedding in exhaled breath and efcacy of face masks. Nature Medicine, 26, 676–680. https://doi.org/
10.1038/s41591-020-0843-2.
Leung, C. C., Lam, T. H., & Cheng, K. K. (2020a). Mass masking in the COVID-19 epidemic: People need guidance. The Lancet, 395, 945. https://doi.org/10.1016/
S0140-6736(20)30520-1.
Leung, G. M., Lam, T. H., Ho, L. M., Ho, S.-Y., Chan, B. H. Y., Wong, I. O. L., & Hedley, A. J. (2003). The impact of community psychological responses on outbreak
control for severe acute respiratory syndrome in Hong Kong. Journal of Epidemiology & Community Health, 57, 857–863.
Liu, F., Qian, H., Zheng, X., Song, J., Cao, G., & Liu, Z. (2019). Evaporation and dispersion of exhaled droplets in stratied environment. IOP Conference Series:
Materials Science and Engineering, 609, Article 042059.
Lloyd-Smith, J. O., Schreiber, S. J., Kopp, P. E., & Getz, W. M. (2005). Super spreading and the effect of individual variation on disease emergence. Nature, 438,
355–359.
Lok, C. (2016). Where sneezes go. Nature, 534, 24–26.
Lo, J. Y., Tsang, T. H., Leung, Y. H., Yeung, E. Y. H., Wu, T., & Lim, W. W. L. (2005). Respiratory infections during SARS outbreak, Hong Kong, 2003. Emerging
Infectious Diseases, 11, 1738–1741.
Loudon, R. G., & Roberts, R. M. (1968). Singing and the dissemination of tuberculosis. American Review of Respiratory Disease, 98, 297–300. https://www.atsjournals.
org/doi/abs/10.1164/arrd.1968.98.2.297.
Marin, A., Karpitschka, S., Noguera-Marín, D., Cabrerizo-Vílchez, M. A., Rossi, M., et al. (2019). Solutal Marangoni ow as the cause of ring stains from drying salty
colloidal drops. Physical Review Fluids, 4, Article 041601.
Marin, A., Liepelt, R., Rossi, M., & K€
ahler, C. J. (2016). Surfactant-driven ow transitions in evaporating droplets. Soft Matter, 12, 1593–1600.
Prather, K. A., Wang, C. C., & Schooley, R. T. (2020). Reducing transmission of SARS-CoV-2. Science. , Article eabc6197. https://doi.org/10.1126/science.abc6197.
Pyankov, O. V., Bodnev, S. A., Pyankova, O. G., & Agranovski, I. E. (2018). Survival of aerosolized coronavirus in the ambient air. Journal of Aerosol Science, 115,
158–163. https://doi.org/10.1016/j.jaerosci.2017.09.009.
Raffel, M., Willert, C. E., Scarano, F., K€
ahler, C. J., Wereley, S. T., & Kompenhans, J. (2018). Particle image Velocimetry. Springer International Publishing AG.
Rensink, D. (2004). Verdunstung akustisch levitierter schwingender Tropfen aus homogenen und heterogenen Medien. Erlangen, Germany: Dissertation.
Reuther, N., & K€
ahler, C. J. (2020). Effect of the intermittency dynamics on single and multipoint statistics of turbulent boundary layers. Journal of Fluid Mechanics,
897. https://doi.org/10.1017/jfm.2020.384. Published online by Cambridge University Press: 10 June 2020.
Rossi, M., Marin, A., & K€
ahler, C. J. (2019). Interfacial ows in sessile evaporating droplets of mineral water. Physical Review E, 100, Article 033103.
Ruedl, G., Kopp, M., & Burtscher, M. (2012). Does risk compensation undo the protection of Ski Helmet use? Epidemiology, 23, 936–937.
C.J. K€
ahler and R. Hain

Journal of Aerosol Science 148 (2020) 105617
12
Sande van der, M., Teunis, P., & Sabel, R. (2008). Professional and home-made face masks reduce exposure to respiratory infections among the general population.
PloS One, 3, Article e2618. https://doi.org/10.1371/journal.pone.0002618.
Scott, M. D., Buller, D. B., Andersen, P. A., Walkosz, B. J., Voeks, J. H., Dignan, M. B., & Cutter, G. R. (2007). Testing the risk compensation hypothesis for safety
helmets in alpine skiing and snowboarding. Injury Prevention, 13, 173–177.
Seto, W. H., Tsang, D., Yung, R. W., Ching, T. Y., Ng, T. K., Ho, M., Ho, L. M., & Peiris, J. S. M. (2003). Effectiveness of precautions against droplets and contact in
prevention of nosocomial transmission of severe acute respiratory syndrome (SARS). The Lancet, 361, 1519–1520.
Soper, G. A. (1919). The lessons of the pandemic. Science, 49, 501–506.
Troyer, E. A., Kohn, J. N., & Hong, S. (2020). Are we facing a crashing wave of neuropsychiatric sequelae of COVID-19? Neuropsychiatric symptoms and potential
immunologic mechanisms. Brain, Behavior, and Immunity, 87, 34–39. https://doi.org/10.1016/j.bbi.2020.04.027. Published on April 13.
Varga, Z., Flammer, A. J., Steiger, P., Haberecker, M., Andermatt, R., Zinkernagel, A. S., Mehra, M. R., Schuepbach, R. A., Ruschitzka, F., & Moch, H. (2020).
Endothelial cell infection and endotheliitis in COVID-19. The Lancet, 395, 1417–1418. https://doi.org/10.1016/S0140-6736(20)30937-5.
Verity, R., Okell, L. C., Dorigatti, I., Winskill, P., Whittaker, C., Imai, N., et al. (2020). Estimates of the severity of coronavirus disease 2019: A model-based analysis.
The Lancet, 20, 669–677. https://doi.org/10.1016/S1473-3099(20)30243-7.
Wang, D., Hu, B., Hu, C., et al. (2020). Clinical characteristics of 138 hospitalized patients with 2019 novel coronavirus-infected pneumonia in Wuhan, China. Journal
of the American Medical Association, 323, 1061–1069. https://doi.org/10.1001/jama.2020.1585.
Wells, W. F., & Wells, M. W. (1936). Air-bone infection (pp. 1698–1703). Cambridge: Journal A. M. A..
Wong, J., Kelly, H., Ip, D. K., Wu, J., Leung, F., & Cowling, B. (2013). Case fatality risk of inuenza A (H1N1pdm09): A systematic review. Epidemiology, 24(6),
830–841. https://doi.org/10.1097/EDE.0b013e3182a67448. Issn print: 1044–3983.
Zandifar, A., & Badrfam, R. (2020). COVID-19: Considering the prevalence of schizophrenia in the coming decades. Psychiatry Research, 288, 112982. https://doi.org/
10.1016/j.psychres.2020.112982.
Zhang, R., Lib, Y., Zhang, A. L., Wang, Y., & Molina, M. J. (2020). Identifying airborne transmission as the dominant route for the spread of COVID-19. PNAS. https://
doi.org/10.1073/pnas.2009637117. First published June 11.
Uvex (2020). https://www.uvex-safety.com/de/wissen/normen-und-richtlinien/atemschutzmasken/die-bedeutung-der-ffp-schutzklassen, called on May 20, 2020.
C.J. K€
ahler and R. Hain