![Areas with comfortable air velocity according to [26]](https://www.researchgate.net/profile/Christian-Kaehler/publication/343514409/figure/fig2/AS:922007172558848@1596835048453/Areas-with-comfortable-air-velocity-according-to-26_Q320.jpg)
Content uploaded by Christian J. Kähler
Author content
All content in this area was uploaded by Christian J. Kähler on Aug 07, 2020
Content may be subject to copyright.
Version from Aug., 7, 2020
1
Can mobile indoor air cleaners effectively reduce an indirect risk of
SARS-CoV-2 infection by aerosols?
Christian J. Kähler, Thomas Fuchs, Rainer Hain
Universität der Bundeswehr München
Institut für Strömungsmechanik und Aerodynamik
Werner-Heisenberg-Weg 39
85577 Neubiberg
Abstract
The worldwide development of the SARS-CoV-2 infection makes it clear that the pandemic is
only just beginning and cannot be stopped. Even if an effective and well-tolerated vaccine were
available, it would not be feasible to vaccinate the world population on a large scale to combat
the spread of the virus. It is therefore necessary to establish technical solutions to contain the
pandemic at least locally. Mouth-and-nose-covers are now generally accepted technical aids
to reduce the direct risk of infection when speaking, singing, coughing and sneezing. However,
indirect infection via infectious aerosols, which accumulate over time in space, cannot be
prevented with simple mouth-and-nose covers or surgical masks [5]. This requires tightly fitting
particle-filtering respiratory masks. Alternatively, the aerosol concentration in the room can be
reduced by means of filtration or discharged via window ventilation. Ventilation systems that
reliably separate aerosol with a diameter smaller than 1 μm are rare. The free ventilation by
means of windows is often not efficient and at the latest in winter no longer possible without
wasting energy and endangering the health and well-being of people. The question is therefore
whether mobile indoor air cleaners are basically suitable for making a meaningful contribution
to reducing the risk of infection? To answer this question, a TROTEC TAC V+ indoor air cleaner
with a volume flow of up to 1500 m3/h was systematically analysed. The unit has a filter
combination that ensures that 99.995% of aerosol with a diameter of 0.1 to 0.3 μm is separated
from the room air. The results show that the aerosol concentration in a room with a size of 80
m2 can be reduced to a low level everywhere within a short time. In our opinion, indoor air
cleaners with a large volume flow and high-quality filters of class H14 represent a very suitable
technical solution to reduce the indirect risk of infection by aerosols in schools, offices, shops,
waiting rooms, community and club houses, lounges and dining rooms, etc. However, they can
also be used as a support in buildings with air conditioning systems where people stand
together (waiting area) and work together or where a lot of aerosol is emitted due to the work
load (fitness studio).
1. Introduction
According to the current state of research, SARS-CoV-2 is mainly transmitted via droplets that
are produced when breathing, speaking, singing, coughing or sneezing and are inhaled and
exhaled through the air we breathe [1, 2, 3, 4]. Direct infection, in which many emitted droplets
Version from Aug., 7, 2020
2
are inhaled over a short distance (less than 1.5 m) by an uninfected person, can be effectively
prevented with the aid of particle-filtering respiratory masks (N95/KN95/FFP2 or better), as
these respiratory masks reliably separate droplets up to a defined size class when inhaled and
exhaled [5]. These respiratory masks therefore offer very good self-protection even without a
safety distance. However, they also have the effect that the viruses emitted by infected persons
hardly reach the air in the room. Therefore, in addition to self-protection, they also offer
protection against direct and indirect infection if no outlet valve is integrated.
Larger droplets are also effectively separated by simple mouth-and-nose covers. Since larger
droplets can statistically transport more viruses than small droplets, this effect is significant for
processes that primarily produce large droplets (coughing, sneezing). The small droplets,
however, are released into the outside air through gaps at the edge of the mouth-and-nose
coverings, because these coverings do not close tightly enough to the face on the one hand
and on the other hand the flow largely follows the path of least flow resistance [5].However,
due to the flow resistance of the mouth-nose-covering and the losses caused by the deflection,
the small droplets remain in the vicinity of the head at first. This greatly limits the short-term
spread of the droplets when speaking, singing, coughing or sneezing and thus the direct risk
of infection is enormously reduced [5, 6, 7, 8, 9]. However, these coverings only provide
effective protection if a safety distance (at least 1.5 m) can be maintained. Since the small
droplets are released by these mouth-and-nose covers, they can accumulate in the room over
time and lead to indirect infection even if the infected person is no longer in the room [8, 10].
Small droplets that float in the air for hours and can be transported over long distances with
the airflow are called aerosol particles and the mixture of air and aerosol particles is called an
aerosol. It is scientifically proven that an aerosol loaded with infectious viruses can lead to
COVID-19 infection as long as the liquid phase of the aerosol particles has not completely
evaporated [11, 12]. The evaporation time of aqueous aerosols with a diameter of a few
micrometers is less than one second at moderate humidity [6]. At high humidity, however, the
evaporation rate can come into equilibrium with the condensation rate and then the liquid
portion is permanently preserved. The probability of infection is then only limited by the half-
life of the viruses. The SARS-CoV-2 half-life is 1.1 - 1.2 hours, i.e. after about an hour,
statistically speaking, half of the viruses in a drop no longer present a risk of infection [11].
Whether aerosol particles are still infectious after evaporation of the liquid phase is currently
the subject of controversial debate [13, 14, 15]. These solid particles consisting of salts, virus
material and dried mucus are called droplet nuclei. Depending on the solid content, the droplet
nuclei are much smaller than the original droplets from which they originated. In chapter 9 we
will go into this in more detail. Due to their small size, the droplet nuclei also float in the air and
therefore, even at low humidity, they form together with the air a long-lasting aerosol that can
accumulate in space over time, provided that sources are present and it is not filtered or
discharged.
Since the indirect probability of infection in a room increases with the number of infected
persons their activity and the length of time the non-infected spend in the room, measures
must be taken to limit the virus load in the room air. Which limit value should be aimed at in
this context is currently not defined and it is also not clear how the SARS-CoV-2 concentration
in the room should be measured at all. Many buildings have ventilation and air conditioning
systems that ensure that contaminated air is removed in a controlled manner and filtered or
fresh air is added from outside. In regions with moderate climatic conditions, however,
Version from Aug., 7, 2020
3
ventilation is usually provided by free ventilation through windows and doors. Recommended
air exchange rates are based on CO2 emissions and other human evaporation (e.g. water
vapour), as well as the accumulation of pollutants in the room (e.g. radon) and the prevention
of building damage (e.g. mould growth) [16]. As early as 1858, Max von Pettenkofer recognized
that indoor CO2 values of a maximum of 1000 ppm should be aimed for [17]. With a CO2
ambient pollution of 450 ppm in cities, however, this value is often difficult to achieve [18]. For
many areas, it is therefore recommended, irrespective of the CO2 concentration, to supply a
certain amount of fresh air per hour, which varies according to the use of the room. In offices,
restaurants and sales rooms, up to 4 ‒ 8 times the room volume should be supplied per hour
in order to protect people and buildings from pollutants and to ensure the performance of
people [18, 19]. If, however, pollutants are emitted into the room that pose a considerable risk
to health, such as viruses, then, depending on the level of risk, considerably higher air
exchange rates of 12 ‒ 15 are necessary [20, 21, 22].
The main advantage of heating, ventilation and air conditioning (HVAC) systems over free
ventilation is that they continuously provide adequate indoor air quality and regular manual
regulation by means of windows is not necessary, see Figure 1 (left). However, they must be
maintained regularly and operated correctly. For energy reasons, they are often operated with
only a small supply of fresh air and simple filters. To prevent indirect infections, however, a
sufficient supply of fresh air or very good filtering of the room air is required [23]. Due to the
mild climatic conditions in many regions, fresh air is often supplied through windows with the
help of free ventilation. With impulse ventilation, the existing pollutant concentration in the room
can be greatly reduced within a short time under suitable wind conditions or temperature
differences, but after closing the windows the pollutant concentration gradually increases
again, as shown schematically in Figure 1 (right).
Figure 1: Schematic representation of the pollutant concentration and the temperature curve
in a room with HVAC system (left) and shock ventilation (right)
In order to keep the pollutant concentration at a low level, it is therefore often necessary to
provide shock ventilation or, better, permanent ventilation. Cross-ventilation is most effective
in suitable weather conditions, but it can also quickly be perceived as unpleasant. A tilted
window, on the other hand, usually does not lead to any significant exchange of air in the room
[16]. Under these conditions, it is recommended to open two windows and to place a fan in
front of one window to bring fresh air into the room from outside. Due to the increase in
pressure, the contaminated air is then discharged through the other window. With this method,
all areas of the room are supplied with fresh air and the dwell time of the aerosol in the room
can be kept short, which reduces the indirect infection risk. In the cold season, however, free
ventilation is not only uncomfortable, but it also endangers health. Furthermore, free ventilation
Version from Aug., 7, 2020
4
is energetically poor and not possible in windowless rooms. Furthermore, the efficiency of air
exchange during free ventilation depends on parameters that cannot be influenced, such as
the position and size of windows, the speed and direction of the wind in front of the windows,
the difference in temperature between inside and outside, and people's willingness to ventilate
[24, 25].
It is difficult to determine which type of ventilation offers the best protection against indirect
SARS-CoV-2 infection, since the number of viruses emitted or the virus concentration and thus
the probability of infection in a room is not known. Since there is also no correlation between
virus concentration and other variables such as CO2 concentration, the virus load cannot be
measured indirectly via the CO2 content. For this reason, the highest possible air exchange
rates should be aimed for in order to minimize the probability of infection. Due to the deadly
danger of a SARS-CoV-2 infection, it is certainly reasonable not to fall below the air exchange
rates already recommended without risk of infection, but to exceed them.
It is foreseeable that in the current pandemic, additional challenges will arise in winter at the
latest in order to minimise the risk of infection in enclosed spaces. However, even the currently
discussed opening of schools without maintaining safety distances and without using mouth-
and-nose covers or particle-filtering respirators is viewed with concern by many people, as
there is a possibility that children will become infected among themselves and the chain of
infection will then run through their parents and grandparents. As COVID-19 mortality
increases with age, this concern of many parents is well founded and justified. It is therefore a
question of whether the viral load can be effectively minimized in buildings without HVAC
systems with class H13/H14 filters or 100% fresh air supply without reducing comfort or
accepting acute damage to health.
It is necessary to find a solution that on the one hand does not cause rapid air movements or
large fluctuations in temperature or humidity in the room, which would impair comfort [26], and
on the other hand efficiently filters out the aerosol particles contaminated with infectious virus
material without distributing them uncontrollably like a "virus slingshot" [4]. To ensure this, a
sufficient air exchange rate is important and the residence time of the emitted aerosols must
be as short as possible. The air exchange rate does not mean that the air is completely
exchanged, as in an air pump, but rather the proportion of fresh or purified air supplied per
hour in relation to the room volume [24]. On the other hand, the air speed and the turbulent
fluctuating movement of the air must not be too great, otherwise it can become unpleasant for
the people in the room, depending on their clothing and activities. Figure 2 illustrates the
dependence of comfort on the average air velocity, the turbulent air fluctuations and the
temperature according to [26]. The areas below the respective curve can be regarded as
comfortable. As a guide value, it is usually required that at moderate room temperatures the
air velocity should be less than 0.3 m/s on average over long periods of time.
The flow movement in space is therefore very important to consider and can be strongly
influenced by fixtures (lamps) or large objects in the room, but also by openings in the walls
(draught) or the movement of people (mixing, thermal convection) [24, 25]. Therefore, the
correct positioning of a room air cleaner is important for efficient filtering of the aerosol particles
in a room with mobile devices. But also the position where the aerosol particles are emitted
locally plays a major role. Further influencing parameters are the emitted aerosol particle
concentration depending on the activity as well as the dwell time in the room.
Version from Aug., 7, 2020
5
Figure 2: Areas with comfortable air velocity according to [26]
Due to the complexity of the flow problem, an experimental approach is used within the scope
of these investigations, since only in this way the many influencing factors can be recorded
physically correctly.
A TROTEC TAC V+ room air cleaner with the following features was used for the experimental
investigations [27]:
1. A maximum volume flow rate of about 1500 m3/h.
2. Class F7 and H14 filters. H14 means that 99.995% of aerosols with a diameter of 0.1
to 0.3 μm are separated. Larger particles are separated to 100%.
3. The filter can be heated to over 100°C after use to destroy the viruses in the filter and
to counteract the formation of biofilms, bacteria and fungi without harmful chemical
additives or UV-C radiation.
4. Despite its weight, the device can be easily moved, so that it can be used in various
areas without any problems.
2. Experimental setup and PIV experiments
The aim of the first series of tests was to quantify the intake and blow-out characteristics of the
indoor air cleaner at different performance levels / volume flow rates. According to Figure 2,
the average air velocity and the turbulent air fluctuations in the environment of the indoor air
cleaner must be quantitatively recorded for this purpose. These variables are important to
determine whether the air movement is perceived as disturbing by people in the vicinity. Air
velocities in the range of less than 0.3 m/s are usually not considered disturbing if the persons
are only engaged in light activities. Since the air movement felt is composed of the average
and turbulent flow movement, the sum of both components must be less than 0.3 m/s on
average. Particle Image Velocimetry (PIV) was used to determine these two quantities in
relation to the distance to the device [28]. In this measurement technique, the movement of
artificially generated aerosol particles, which follow the air movement exactly, is registered in
a laser light-sheet with digital cameras at two instants in time and then the displacement of the
particle images is determined using digital image processing methods. From the displacement
Version from Aug., 7, 2020
6
of the particle images, the spatially resolved velocity distribution in the measuring plane can
then be determined without contact, taking into account the time interval between the
measurements and a calibration. Figure 3 shows pictures of the test setup and during a
measurement.
Figure 3: Experimental setup with indoor air cleaner, two double pulse lasers and 4 sCMOS
cameras (left) and recording during measurement with PIV (right)
For the experiments 4 PCO.edge 5.5 sCMOS cameras with Zeiss macro lenses with focal
lengths of 35 mm and 50 mm were used. The aerosol particles were generated from di-2-
ethylhexyl sebacate (DEHS) with a seeding generator from the company PIVTEC. The
average diameter of the aerosol particles is 1 μm and the size distribution is between 0.1 ‒ 2
μm [29]. For the PIV flow experiments in chapter 6 helium filled bubbles with an average
diameter of about 350 μm were used, which were generated with an HFSB generator from
LaVision. Two Quantel Evergreen 200 lasers were used to illuminate the particles, whose
beams were formed into light-sheets using different lenses [28]. The measuring system was
controlled by the software DaVis from LaVision GmbH, which was also used for data
evaluation.
3. Characterisation of the flow field in the vicinity of the air purifier
Figures 4 ‒ 9 show the results of measurements at volume flows of 600 m3/h, 1000 m3/h and
1500 m3/h. In the upper figure, the velocity magnitude is colour-coded and the direction of the
average flow velocity is shown as vectors. In the lower figure, the amount of turbulent air
movement is shown. The quantitative measurement results show that the mean flow velocity
and the turbulent motion in the vicinity of the device do not exceed the critical values shown in
Figure 2. Therefore, even if you stay in the immediate vicinity of the air purifier, no impairment
of your well-being by air movements is to be expected. At a volume flow of 1500 m3/h, slightly
higher flow speeds are achieved in front of the intake area, but only in the leg area and only
up to a distance of approx. 0.5 m in front of the unit. In the sensitive head and body area, the
average flow velocities and turbulent air movement are significantly lower than 0.3 m/s. Directly
in the outflow area and above the head of the device, considerably higher flow speeds are
achieved. At a distance of 0.5 m from the unit, however, these are not noticeable even when
the unit is standing, as the emerging free jet is directed strongly upwards by baffle plates. It is
therefore hardly possible to block the outlet jet, which could lead to an impairment of the filter
effect in the room. As a result, it can be stated that the air velocities comply with the guidelines
due to the operation of the room air cleaner.
Version from Aug., 7, 2020
7
Figure 4: Average flow field measured at 600 m3/h.
Figure 5: Turbulent flow movement averaged over time at 600 m3/h.
Version from Aug., 7, 2020
8
Figure 6: Average flow field measured at 1000 m3/h.
Figure 7: Turbulent flow movement averaged over time at 1000 m3/h.
Version from Aug., 7, 2020
9
Figure 8: Average flow field measured at 1500 m3/h.
Figure 9: Turbulent flow movement averaged over time at 1500 m3/h.
Version from Aug., 7, 2020
10
4. Flow effects through objects
An independent study has already shown that the filter effect can depend on the location of
the device and objects in the room [30]. To determine the extent to which positioning the room
air purifier close to a wall and long ceiling lamps can adversely affect the flow movement in the
room and thus possibly delay the time for filtering the room air, the room air purifier was
positioned close to a wall and the outlet jet was blocked by room-dividing ceiling lamps. The
velocity distribution shown in Figure 10 illustrates that the discharged air first rises to the ceiling
and is then deflected at the ceiling. Due to the Coandă effect, the outlet jet is following the
ceiling and spreads out as a so-called wall jet. Due to the protruding ceiling lamps, which run
along the entire length of the room, the flow is again deflected and consequently a large vortex
is formed. These local vortices reduce the filtering performance in the room, as the filtered air
is partly returned to the unit and filtered again before sufficient entrainment of viruses in the
room can take place [24]. Such constellations should therefore be avoided as far as possible
by suitable positioning of the unit in the room.
Figure 10: Measured room air flow in front of the room air cleaner with interrupted ceiling
In order to enable effective filtering of the room air, it is recommended that the room air cleaner
is positioned in the middle of the longest side of the room if possible, and that it is ensured that
the air jets hitting the ceiling can flow along the ceiling undisturbed for as long as possible. If
the ceiling is smooth and the room is not too large, the jet can flow to the other side wall. There
it is deflected downwards and flows along the floor back to the intake area of the purifier, as
Version from Aug., 7, 2020
11
shown schematically in figure 11 (left). In this way, the best possible air exchange is ensured,
as the contaminated air areas in the room are continuously transported to the unit via a few
large vortex flows. If the air flow on the ceiling is interrupted, e.g. by a lamp, a flow movement
as shown in Figure 11 (right) can be expected, provided the room is sufficiently large.
Figure 11: Simplified schematic representation of the room air flow in an empty room (left)
and in the presence of a row of lamps on the ceiling (right). In reality, the flow phenomena
are three-dimensional.
The obstruction of the air flow at the ceiling can also be caused by a high heat source (heating,
stove, group of people) in the room. Heat sources cause the density of the air to decrease and
due to the reduced density it becomes lighter than the ambient air and therefore it is moved
upwards by the heavier air masses at the bottom and rises. This thermal convection flow then
reaches the ceiling where it is deflected in all directions. This flow can therefore counteract the
flow movement generated by the fan and lead to the flow separation of the wall jet at the ceiling.
In this case, a vortex can again form which does not cover the entire room, and the filter
performance can thus be reduced, analogous to the schematic representation in Figure 11
(right). It should be noted, however, that the turbulent air movement in the room will always
ensure that all the air in the room enters the fan's area of influence after a certain time and the
aerosol is then separated. However, the aim should be to keep the time for filtering the room
air as short as possible.
5. Experimental design and performance of concentration measurements
With the aim of quantitatively determining the filter performance of the room air purifier,
concentration measurements were carried out simultaneously at six positions in the 80 m2
room. The particle imaging method was used to determine the decay rates of the aerosol
concentration in the room air. For this purpose, the room was first nebulized homogeneously
and at high concentration with very long-lived aerosol particles whose size distribution
corresponds to the aerosol emitted when breathing, speaking and singing. The longevity of the
aerosol particles is very important, as otherwise the measurement results would be falsified by
evaporation. Furthermore, the small size is important because otherwise the aerosol particles
would settle over time, which would also cause systematic measurement errors. The aerosol
particles are illuminated by a pulsed laser and imaged by a camera with a suitable lens and
stored for further processing. The number of particle images on the sensor corresponds to the
number of aerosol particles in the illuminated volume. The number on the sensor must not be
too high as overlapping particle images would systematically falsify the count. For this reason,
the imaging scale of the optical system must be adapted to the initial concentration of the
aerosol particles. To detect the particle images, digital filters are used to suppress the noise,
and then the background is removed by means of an intensity threshold. As a result of this
Version from Aug., 7, 2020
12
image pre-processing, only the aerosol particle images remain on the image, which are then
counted. Without this image pre-processing, stochastic image noise could be misinterpreted
as a signal, which would lead to systematic measurement errors. By taking images at a fixed
frequency over a sufficiently long period of time, the individual aerosol particle images can be
reliably counted in each individual image. The result of the measurements is the number in the
measurement volume over time, from which parameters such as the decay rate, half-life etc.
can be determined.
To analyse the functionality of the TROTEC TAC V+ room air purifier in an 80 square metre
room, the aerosol particle concentration was measured simultaneously at 6 independent
locations in the room. For reasons of symmetry, the concentration was only determined on the
axis of symmetry of the room and in one half of the room. With a PIVTEC seeding generator
the aerosol particles with a size distribution between 0.1 ‒ 2 μm and a mean diameter of about
1 μm was generated from Di-2-ethylhexyl-sebacate (DEHS) [19]. For recording the aerosol 6
PCO.edge 5.5 sCMOS cameras with Zeiss macro lenses with a focal length of 50 mm were
used. For the illumination of the aerosol particles in the measuring volume, 2 Quantel
Evergreen 200 lasers were used, whereby the lasers were set up in such a way that 3
measuring points (cameras) each share a laser beam to illuminate the local measuring volume
(MP1 ‒ MP3 and MP4 ‒ MP6). The individual cameras and lasers were centrally controlled
by LaVision's DAVIS software so that the recordings of all 6 cameras were performed
synchronously. The recording rate was 1 Hz.
Figure 12 shows the optically distorted panoramic image of the test setup in the room and
Figure 13 shows the geometry and dimensions of the room as well as the positions of the
device and the six measuring points in top view. Two instrument positions were chosen.
Position A can be regarded as a quasi optimal position. A position in the middle of the room
would be even more advantageous with regard to the filter efficiency according to Figure 11,
but a central installation position will rarely be realized in practice. Position B indicates the most
unfavourable position in the room, as the installation in the corner of the room and the lamps
on the ceiling will prevent an optimal indoor air flow, according to Figure 10.
Figure 12: Optically distorted panoramic image of the test chamber with the components for
the concentration measurements
Version from Aug., 7, 2020
13
Figure 13: Arrangement of the components in the room for concentration measurements
6. Results of the concentration measurements
Using the test set-up used in section 2, the aerosol particle separation efficiency at the outlet
of the unit was first verified visually. For this purpose, the entire room was homogeneously
nebulized with DEHS aerosol particles with a diameter of 0.1 ‒ 2 μm and an average diameter
of approx. 1 μm [19]. With the help of a laser light-sheet in the area of the outflow it was
measured whether aerosol was still coming out of the outlet area of the unit. Figure 14 shows
that the free jet emerging at the upper corner of the room air purifier is free of aerosol at the
different volume flows (black area). The performance of the Class H14 filter is therefore clearly
visible even at high volume flows. The turbulent structure of the free jet and the entrainment,
i.e. the mixing of aerosol into the clean free jet, is also clearly visible [24]. This entrainment or
mixing effect is very important for an efficient transport of the aerosol to the intake area of the
room air cleaner. The area surrounding the free jet, on the other hand, is completely
contaminated with aerosol particles (white areas). The high concentration outside the free jet
illustrates the average aerosol particle concentration in the room generated for the
measurements.
Version from Aug., 7, 2020
14
Figure 14: Representation of the aerosol distribution in the outflow area at a volume flow of
1000 m3/h
Figure 15 shows an example of the course of the aerosol particle concentration over time. The
left Figure shows the concentration with the room air purifier switched off as a reference. Since
all openings in the room have been sealed tightly and the aerosol particles hardly settles at all,
the concentration is largely constant over the considered measuring time, so that an influence
of natural deposition processes on the concentration measurements during measurements
with the device switched on is not affected. In addition, tests were carried out with the unit
running without F7 and H14 filters. The fact that the decay in particle concentration is slightly
greater than in the case without operation of the room air cleaner is due to the fact that a very
small amount of aerosol particles aer separated in the fan due to centrifugal forces. However,
a liquid film could not be detected on the fan or in the housing of the blower.
Figure 15: Decrease of aerosol concentration over time for different volume flows and
associated exponential fit functions
The results clearly show that small long-lived droplet nuclei or droplets remain in the air almost
permanently under conditions where the evaporation rate is in equilibrium with the
condensation rate. The values given for the decay times in Figure 15 (left) make it possible to
estimate how long it takes for the concentration to reach a desired value due to settling or
rejection phenomena at walls. Due to the very small decay constant it is clear that the aerosol
Version from Aug., 7, 2020
15
can cause infections for many hours if it is not filtered out or if the temporal deactivation of the
viruses makes an infection unlikely.
It is also obvious that the aerosol particles, since they do not evaporate completely, can be
transported over very long distances (in principle many kilometres) by the air flow. However, it
must be taken into account that during this transport the concentration and thus the probability
of infection decreases enormously due to two processes. On the one hand, turbulent diffusion
ensures spatial dispersion of the aerosol, which also occurs when the average flow velocity is
zero. On the other hand, the aerosol released over a period of time is spatially strongly
stretched and therefore diluted if the average flow velocity is not zero. For example, if 1 litre of
air is exhaled over a period of 2 seconds during light physical exertion and the surrounding air
has a velocity of 1 m/s, then the exhaled aerosol is stretched over a range of 2 m due to the
flow. The concentration will therefore mathematically decrease by a factor of 20 and thus the
viral load in the wake of the person from whom the aerosol is exhaled. For example, if viruses
are exhaled by a cyclist riding at a speed of 10 m/s, the virus load in the wake will be diluted
by a factor of 200 based on the speed alone. If turbulent diffusion is also taken into account,
the concentration decreases by at least 1-2 orders of magnitude. In addition, not all aerosol
particles carry viruses, see chapter 8. In the open air, an aerosol infection is therefore
extremely unlikely if the wind speed is sufficient or the person moves.
Figure 15 (right) shows the decrease of the aerosol concentration in the 80 m2 room as a
function of time and the volume flow of the room air cleaner. The very efficient decrease of the
concentration within a few minutes shows the efficiency of the F7 / H14 filter combination in
connection with the volume flow of the room air cleaner. The exponentially decreasing course
of the aerosol concentration makes it possible to quantitatively determine characteristic
quantities which are essential for the evaluation of the filter performance. The decay constant
is a measure for the efficiency of the filtration. The higher the value, the better the filter effect
and the shorter the time required to filter the room air. The half life indicates how long it takes
for the aerosol concentration at the point of measurement to decrease to half. The mean
residence time characterizes how long the aerosol emitted at the respective measuring
Table 1: Decrease in aerosol concentration over time for different volume flows. Decay rate
(black), half-life (green) and mean residence time (blue).
Measuring
position 600 m3/h
(Position A) 1000 m3/h
(Position A) 1500 m3/h
(Position A) 1000 m3/h
(Position B)
MP1 3,30 | 0,21 | 0,30 4,13 | 0,17 | 0,24 6,06 | 0,11 | 0,17 3,35 | 0,21 | 0,30
MP2 3,04 | 0,23 | 0,33 3,86 | 0,18 | 0,26 5,95 | 0,12 | 0,17 3,04 | 0,23 | 0,33
MP3 2,90 | 0,24 | 0,34 3,48 | 0,20 | 0,29 5,48 | 0,13 | 0,18 3,22 | 0,22 | 0,33
MP4 3,11 | 0,22 | 0,32 4,03 | 0,17 | 0,25 6,16 | 0,11 | 0,16 3,47 | 0,20 | 0,29
MP5 3,05 | 0,23 | 0,33 3,82 | 0,18 | 0,26 6,05 | 0,11 | 0,17 3,25 | 0,21 | 0,31
MP6 2,86 | 0,24 | 0,35 3,65 | 0,19 | 0,27 5,62 | 0,12 | 0,18 2,95 | 0,24 | 0,34
Decay rate [1/h] | Half-life [h] | Mean residence time [h]
The filter performance is strongly dependent on the volume flow of the room air cleaner, but
despite the size of the room, only slightly dependent on the distance to the unit. At 600 m3/h,
Version from Aug., 7, 2020
16
the aerosol concentration is halved at position 1 after about 12 minutes and at the furthest
position 6 after about 14 minutes. At a volume flow of 1500 m3/h, halving the aerosol
concentration at position 1 takes about 6 minutes and 7 minutes at position 6.
For comparison, the right-hand column shows the values for a volume flow of 1000 m3/h for
position B of the room. In comparison with the results at position A for the same volume flow,
there is an impairment of the filter performance of about 20%. A poor position can therefore be
compensated for by a higher volume flow, but then the energy requirement is greater and the
noise development higher for the same filter performance. Therefore, the choice of position is
definitely an important point.
In Table 2, a loaded and a new filter of Class H14 were measured comparatively. Taking into
account the measurement uncertainty of 2σ=0.14 [1/h], no significant difference can be found
between the two filters. It is therefore to be expected that the filters can have a service life of
several years, depending on the load.
Table 2: Comparison between a loaded (used) and an unloaded (new) H14 filter
Measuring
position 1000 m3/h (Position
A), used filter 1000 m3/h (Position
A), new filter
MP1 4,13 | 0,17 | 0,24 4,22 | 0,16 | 0,24
MP2 3,86 | 0,18 | 0,26 3,75 | 0,18 | 0,27
MP3 3,48 | 0,20 | 0,29 3,50 | 0,20 | 0,29
MP4 4,03 | 0,17 | 0,25 4,09 | 0,17 | 0,24
MP5 3,82 | 0,18 | 0,26 3,82 | 0,18 | 0,26
MP6 3,65 | 0,19 | 0,27 3,60 | 0,19 | 0,28
Decay rate [1/h] | Half-life [h] | Mean residence time [h]
7. Dependence of the filter performance on the room geometry
The filter performance not only depends on the unit and the installation site, but also on the
geometry of the room. Long rooms in particular are more difficult to filter because the jet on
the ceiling eventually detaches and then a recirculation area forms that does not reach the
opposite wall. This situation is comparable to the situation in Figure 11 (right) where the
detachment of the flow from the ceiling is not caused by an object but by the reduction of the
impulse of the wall jet with increasing distance. The reduction of the impulse is caused by the
wall friction, the turbulent air movement and the entrainment, which can be seen very clearly
in Figure 14. Especially the entrainment accelerates slow flow areas with aerosol through the
filtered wall jet and the work required for this leads to a reduction of the wall jet momentum.
The turbulence primarily leads to a widening of the jet, which also leads to a local decrease of
the momentum and therefore shifts the flow-separation position of the wall jet closer to the unit.
Due to these effects it is possible that the front area of the room is filtered very well, but the
rear area is not. To investigate this situation generically, measurements were made in an
elongated room with a cross-sectional area of approx. 4 m2 . Two different room lengths were
investigated: 22.4 m (experimental setup see Figure 16) and 11.8 m (Figure 17).
Version from Aug., 7, 2020
17
Figure 16: Arrangement of the components in the long corridor configuration for
concentration measurements
Figure 17: Arrangement of the components in the short corridor configuration for
concentration measurements
The quantities determined from the concentration measurements are shown in Tables 3 and
4. It can be clearly seen that even in these elongated rooms a fairly fast separation of the
aerosol is achieved. Even at the far measuring point MP2 in the long configuration a clear
aerosol decrease can be observed, which corresponds approximately to the decrease at
position MP1. As expected, the aerosol is removed faster in the short corridor configuration. It
follows from this analysis that for very long rooms the use of two room air cleaners at the
respective ends may be recommended.
Table 3: Decrease of aerosol concentration in the long corridor configuration over time for
different volume flows. Decay rate (black), half-life (green) and mean residence time (blue).
Volume flow MP1 MP2
600 m³/h 4,21 | 0,16 | 0,24 3,98 | 0,17 | 0,25
1000 m³/h 4,83 | 0,14 | 0,21 4,49 | 0,15 | 0,22
1585 m³/h 6,98 | 0,10 | 0,14 7,33 | 0,09 | 0,14
Decay rate [1/h] | Half-life [h] | Mean residence time [h]
Version from Aug., 7, 2020
18
Table 4: Decrease in aerosol concentration in the short corridor configuration over time for
different volume flows. Decay rate (black), half-life (green) and mean residence time (blue).
Volume flow MP1
600 m³/h 7,37 | 0,09 | 0,14
1000 m³/h 9,91 | 0,07 | 0,10
1585 m³/h 14,57 | 0,05 | 0,07
Decay rate [1/h] | Half-life [h] | Mean residence time [h]
8. Comparison of the separation rate of the class F7 and H14 filter
HVAC systems for fine dust separation with high air purity are usually not equipped with class
H14 filters, but often only with class F7 filters. The question therefore arises as to whether a
class F7 filter is sufficient to reliably and efficiently separate aerosol generated when breathing,
speaking, singing, coughing and sneezing. In [23] it is claimed that with double filtration of the
room air with a class F7 filter, a total of 99% of airborne bacteria and viruses are removed from
an air stream. SARS-CoV-2 is about 0.15 μm, but since it is transmitted either by droplets or
droplet nuclei, separation efficiencies in the range 0.3 ‒ 1 μm should be considered. The
collection efficiency of a class H14 filter is 99.995% for particles with a diameter in the range
of less than 0.3 μm and almost 100% for all larger diameters. For a class F7 filter, it is 40 ‒
65% in the range 0.3 ‒ 1 μm and even less in the range smaller than 0.3 μm. Even a tenfold
filtering of the aerosol with a class F7 filter gives a worse separation result than a simple filtering
with a class H14 filter. The assumption that a large volume flow with an F7 filter leads to
comparable results as an H14 filter with a small volume flow is therefore not plausible for the
size class under consideration. It must also be taken into account that the separation of small
particles is based on physical mechanisms that only work efficiently if the flow velocity in the
filter is low. A significant increase in the volume flow rate is therefore only advisable if the total
surface area of the filter is also increased. A high volume flow inevitably also leads to a high
air velocity in the room and this can be unpleasant according to Figure 2. Furthermore, noise
also increases with the air speed. Finally, it must be taken into account that a high air velocity
in the ventilation ducts leads to an increase in losses and thus the energy requirement
increases. On the other hand, the flow resistance of a class F7 filter is significantly lower than
that of a class H14 filter, which speaks in favour of the class F7 filter. In order to be able to
compare the filter performance of the two filters, tests were carried out at a volume flow rate
of 1000 m3/h in the room shown in Figure 17.
The results of the investigation show that the half-life for the F7 + H14 configuration is 4.2
minutes, for the F7 filter alone 10.2 minutes and 64 minutes without filter, see Table 5. The
analysis shows that an HVAC system with class F7 filters achieves good separation
efficiencies with multiple filtration and therefore an existing HVAC system should be operated
with class F7 or better filters. Since aerosol particles with a diameter in the submicron range
cannot be efficiently removed with the class F7 filter, the fresh air portion should be selected
as large as possible to compensate for the considerable disadvantage compared to the class
H14 filter in the separation of the small aerosol particles.
Version from Aug., 7, 2020
19
If fast and highly efficient filtering is required that reliably filters out even the smallest aerosol
particles with high efficiency and small volume flow, then a class H14 filter is recommended.
In areas that do not have air handling units and are not too voluminous, mobile room air filters
offer a very good possibility to remove the aerosol in the room without the disadvantages of
free ventilation.
Table 5: Decrease in aerosol concentration in the short corridor configuration over time for
different filter configurations. F7 + H14, only F7, no filters, device switched off. Decay rate
(black), half-life (green) and mean residence time (blue).
Volume flow MP1
1000 m³/h (F7 + H14) 9,91 | 0,07 | 0,10
1000 m³/h (only F7) 4,15 | 0,17 | 0,24
1000 m³/h (no filter) 0,65 | 1,07 | 1,54
0 m³/h 0,66 | 1,05 | 1,52
Decay rate [1/h] | Half-life [h] | Mean residence time [h]
9. Consideration of virus-laden particles and their infectivity
Within the scope of this study, primarily the separation of aerosol particles with a diameter in
the range of 0.3 - 2 μm was considered, since this size is physically difficult to separate on the
one hand and because on the other hand this size is particularly relevant for SARS-CoV-2
infections from our point of view. When considering the aerosols potentially considered
dangerous, 2 facts must be considered. As explained at the beginning, a distinction must be
made between the wet and dry state of the aerosol particles. If the particles are produced in
the lungs or the respiratory tract, they are initially wet, i.e. the solids (salts, proteins, ...) and
any viruses present are in an aqueous environment. If these wet aerosol particles leave the
body, the aqueous phase evaporates in a short time when the air humidity is moderate. For
small particles with diameters in the range of a few micrometers, this happens within fractions
of a second at low ambient humidity [6]. What remains are solid droplet nuclei with only a very
low water content. If we assume for simplicity's sake that all the water evaporates, the
diameters of the dry aerosol particles shown in Figure 18 are based on the wet aerosol
particles. Three curves are shown for different typical solid mass fractions of 1, 3 and 5% [31].
The density of water was assumed to be 1000 kg/m³, the density of the solids 1300 kg/m³.
The diameter of the aerosol particles released is usually in the range of less than 10 μm, with
the maximum of the size distribution being approximately 1 μm [32, 33]. For a direct infection,
diameters in this range are relevant, but for an indirect infection, the diameters readable from
Figure 18 must be taken into account for the respective solid contents. The latter are also
relevant for filtering at moderate humidity. The size of the aerosol that must be separated is
therefore in the range 0.3 ‒ 3 μm according to Figure 18, whereby according to this calculation
sizes around 0.3 μm make up the majority of the aerosol after evaporation. According to this
consideration, the aerosol size distribution primarily considered by us in this study is of high
relevance with regard to SARS-CoV-2 infection if the droplet nuclei are found to be infectious.
Version from Aug., 7, 2020
20
Figure 18: Diameter of dried aerosols as a function of the initial diameter of wet aerosols for
3 different solid mass fractions
Furthermore, the question arises how high the probability is that an aerosol particle contains a
virus at all. For this purpose, a simple estimation was performed, whereby the virus
concentration [viruses/ml] was related to the number of aerosol particles per volume
[droplets/ml] at a certain aerosol diameter. The result of this consideration is shown in Figure
19. The probabilities are given for 4 different virus concentrations resulting from [34]. According
to these investigations, 7 × 106 viruses/ml were an average virus concentration and 2.35 × 109
viruses/ml were the maximum virus concentration in the respiratory mucosa. It can be seen
that especially small aerosols have only a low probability of transporting a virus. Not every
aerosol particle released by an infected person is therefore infectious.
Figure 19: Result of a simple estimation of the probability that a wet aerosol contains a virus
Version from Aug., 7, 2020
21
If it is further assumed that a dry aerosol has a diameter of 1 µm, Figure 18 shows that with a
solid content of 1%, the diameter of the wet particle is 5 µm. At a virus concentration in the
respiratory mucosa of 1 × 109 viruses/ml, this results in a probability that this aerosol contains
a virus of about 7%. Thus, only about every 14th aerosol particle is infectious. On the other
hand, it is precisely the small aerosols in the room air that predominate by far, so that they
cannot be neglected, especially if people spend a long time in a room contaminated with
infectious aerosol particles. Therefore a very good and efficient separation of the small
diameter aerosol or a good supply of fresh air to dilute the viral load in the room is very
important to minimize the indirect risk of SARS-CoV-2 infection.
10. Summary and conclusion
The quantitative measurement results show that with the tested TROTEC TAC V+ room air
cleaner, due to the large volume flow and the filter combination of class F7 + H14, the aerosol
particle concentrations can be halved in 6 ‒ 15 minutes, depending on the volume flow, in
rooms with a surface area of 80 m2 . In rooms with a surface area of 20 m2, halving is achieved
in 3 ‒ 5 minutes, depending on the volume flow rate. It is therefore possible with room air
cleaners to keep the aerosol concentration in small and medium-sized rooms at a low level
without any problems.
Even in a 22 m long, corridor-like room with more than 40 m2, a halving of the aerosol
concentration could be realised within about 5 minutes at maximum volume flow. In larger
rooms, rooms with many objects or very unfavourable geometries, several air purifiers should
be used if necessary in order to filter all areas of the room quickly. Due to the dangerousness
of the SARS-CoV-2 infection, the air exchange rate should, in our opinion, reach at least values
in the range 4 - 8.
To enable the most effective filtering of the room air, the room air cleaner should be positioned
in the middle of the longest side of the room if possible. In addition, the ceiling area should not
be interrupted by objects in the direction of the outflows, if possible, as otherwise the spread
of the wall jets is disturbed and an unfavourable vortex can develop in the room. If operating
conditions are unfavourable, the volume flow rate should be increased to ensure adequate
filter performance. It is also recommended that the unit be operated in continuous operation
and not intermittently, so that no increased virus concentration can develop in the room.
Powerful room air cleaners with class F7 + H14 filter combination can keep the aerosol
concentration in small and medium-sized rooms at a low level and therefore the indirect risk of
infection can be greatly reduced by these units, even with closed windows and without a
suitable air conditioning system. They are therefore very well suited to permanently ensure a
low viral load in classrooms, shops, waiting or treatment rooms, for example, without having
to worry about opening windows and impairing the well-being in the room. Furthermore, in
contrast to free ventilation with windows, they also ensure that there is a real reduction in the
virus load, which often cannot be guaranteed with free ventilation. They also offer the
advantage over HVAC systems that are operated with little or no fresh air, that the viruses are
really eliminated and are not distributed via other channels in the building.
For hygienic reasons, the tested room air cleaner is able to decontaminate the H14 filter by
heating it to about 100°C. If this happens daily for about 30 minutes, the viruses in the filter are
deactivated and the formation of biofilms, bacteria and fungi can be prevented without harmful
chemical additives or UV-C radiation.
Version from Aug., 7, 2020
22
Finally, it should be emphasized that although room air purifiers, open windows and powerful
air handling units are suitable tools to counteract the indirect risk of infection, they cannot
reduce the direct risk of infection, which can occur through direct coughing or during long
conversations over short distances. It is therefore important to maintain sufficient distance from
other people and to wear mouth-and-nose masks or particle-filtering respirators to prevent
direct infection.
Note
The investigations were financially supported by the company TROTEC GmbH, Heinsberg,
Germany. The TAC V+ room air cleaner was provided by TROTEC for the investigations. The
investigations were carried out in accordance with good scientific practice. The support
provided by TROTEC has no effect on the results presented.
Literature
[1] Jayaweera M, Perera H, Gunawardana B, Manatunge J (2020) Transmission of COVID-
19 virus by droplets and aerosols: A critical review on the unresolved dichotomy. Environ
Res. 2020 Sep; 188: 109819. https://doi.org/10.1016/j.envres.2020.109819
[2] Morawska L, Cao J (2020) Airborne transmission of SARS-CoV-2: The world should face
the reality. Environment International, Volume 139, 105730
[3] Zhang R, Lib Y, Zhang AL, Wang Y, Molina MY (2020) Identifying airborne transmission
as the dominant route for the spread of COVID-19. PNAS (2020),
10.1073/pnas.2009637117, first published June 11
[4] Lu J et al. (2020) COVID-19 Outbreak associated with air conditioning in restaurant,
Guangzhou, China, 2020. Emerging Infectious Diseases, www.cdc.gov/eid, Vol. 26, No.
7, DOI: https://doi.org/10.3201/eid2607.200764
[5] Kähler CJ, Hain R (2020) Fundamental protective mechanisms of face masks against
droplet infections. Journal of Aerosol Science 148:105617,
https://doi.org/10.1016/j.jaerosci.2020.105617
[6] Kähler CJ, Hain R (2020) Strömungsanalysen zur SARS-CoV-2 Schutzmaskendebatte.
https://www.unibw.de/lrt7/stroemungsanalysen-zur-sars-cov-2-schutzmaskendebatte
https://youtu.be/SM2QrPFC3MY
[7] Kähler CJ, Hain R (2020) Musizieren während der Pandemie - was rät die Wissenschaft?
https://www.unibw.de/lrt7/musizieren-waehrend-der-pandemie
https://youtu.be/0JmcjRhV-rs
[8] Wells WF, M.W. Wells MW (1936) Air-bone infection. Journal A. M. A., Cambridge, pp.
1698‒1703
[9] Soper GA (1919) The lessons of the pandemic. Science, 49, pp. 501‒506
[10] Fernstrom A, Goldblatt M (2013) Aerobiology and its role in the transmission of infectious
diseases. Journal of Pathogens, Volume 2013, article ID 493960,
http://dx.doi.org/10.1155/2013/493960
[11] Doremalen van N, Bushmaker T, Morris DH, Holbrook MG, Gamble A, Williamson BN,
Tamin A, Harcourt JL, Thornburg NJ, Gerber SI, Lloyd-Smith JO, de Wit E, Munster VJ
Version from Aug., 7, 2020
23
(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
[12] Pyankov OV, Bodnev SA, Pyankova OG, Agranovski IE (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
[13] Morawska L, Milton DK (2020) It is Time to Address Airborne Transmission of COVID-
19. Clinical Infectious Diseases, ciaa939, https://doi.org/10.1093/cid/ciaa939
[14] Klompas M, Baker MA, Rhee C (2020) Airborne Transmission of SARS-CoV-2 -
Theoretical Considerations and Available Evidence. JAMA. Published online July 13,
doi:10.1001/jama.2020.12458
[15] McKie R (2020) Did singing together spread coronavirus to four choirs? 17. May, The
Guardian / The Observer, https://www.theguardian.com/world/2020/may/17/did-singing-
together-spread-coronavirus-to-four-choirs
[16] Heinz E (2016) Wohnraumlüftung – frei und ventilatorgestützt. Beuth Verlag GmbH,
Berlin, Wien, Zürich
[17] von Pettenkofer M (1858) Besprechung allgemeiner auf die Ventilation bezüglicher
Fragen. Über den Luftwechsel in Wohngebäuden. J.G. Cottaische Buchhandlung,
München
[18] DIN EN 16798 – Gebäude-Energieeffizienz mit LüKK-Systemen
[19] VDI 2082 – Raumlufttechnik Verkaufsstätten (VDI-Lüftungsregeln)
[20] DIN 1946-4 – Raumlufttechnik im Gesundheitswesen
[21] Deutsche Gesellschaft für Krankenhaushygiene (2002) Leitlinienentwurf: Ausführung
und Betrieb von raumlufttechnischen Anlagen (RLT-Anlagen) in Krankenhäusern. Hyg
Med 27. Jahrgang 2002 – Heft 3
[22] Deutsche Gesellschaft für Krankenhaushygiene (2015) Krankenhaushygienische
Leitlinie für die Planung, Ausführung und den Betrieb von Raumlufttechnischen Anlagen
in Räumen des Gesundheitswesens. Hyg Med 2015; 40 – 12
[23] Fachverband Gebäude-Klima e. V. (2020) Betrieb Raumlufttechnischer Anlagen unter
den Randbedingungen der aktuellen Covid-19-Pandemie. 03.08.2020, Version 3,
www.fgk.de/images/Aktuelles/2020/08-20/RLT_Covid19_V3_200803.pdf
[24] Etheridge D, Sandberg M (1996) Building Ventilation: Theory and Measurement. Wiley
ISBN: 978-0-471-96087-4
[25] Etheridge D (2011) Natural Ventilation of Buildings: Theory, Measurement and Design.
Wiley ISBN: 978-0-470-66035-5
[26] DIN 1946 Teil 2 (1994) Raumlufttechnik: Gesundheitstechnische Anforderungen (VDI-
Lüftungsregeln)
[27] https://de.trotec.com/produkte-services/maschinen-
highperformance/luftreinigung/raumluftreiniger-tac-v/
[28] Raffel M, Willert CE, Scarano F, Kähler CJ, Wereley ST, Kompenhans J (2018) Particle
Image Velocimetry: a practical guide. Springer, 3. Auflage, ISBN-10: 9783319688510
Version from Aug., 7, 2020
24
[29] Kähler CJ, Sammler B, Kompenhans J (2002) Generation and control of tracer particles
for optical flow investigations in air. Experiments in Fluids 33:736–742
[30] Küpper M, Asbach C, Schneiderwind U, Finger H, Spiegelhoff D, Schumacher S (2019)
Testing of an indoor air cleaner for particulate pollutants under realistic conditions in an
office room. Aerosol and Air Quality Research 19:1655–1665, doi:
10.4209/aaqr.2019.01.0029
[31] Stadnytskyi V, et al. (2020) The airborne lifetime of small speech droplets and their
potential importance in SARS-CoV-2 transmission. PNAS 117:11875–11877
[32] 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, pp. 840–853
[33] Johnson GR, Morawska L, Ristovski, ZD, et al. (2011) Modality of human exposed
aerosol size distributions. Journal of Aerosol Science 42:839–851
[34] Wölfel R, et al. (2020) Virological assessment of hospitalized patients with COVID-2019.
Nature 581:465–469
