Content uploaded by Paolo Arguelles
Author content
All content in this area was uploaded by Paolo Arguelles on Apr 28, 2020
Content may be subject to copyright.
arXiv:submit/3147435 [q-bio.QM] 25 Apr 2020
1
Estimating UV-C Sterilization Dosage for
COVID-19 Pandemic Mitigation Efforts
Paolo Arguelles
pa394@cornell.edu
Abstract—The COVID-19 pandemic caused by SARS-CoV-
2hasinvokedwidespreadinterestineffectiveandreliable
disinfection methods to help combat the virus, including ultra-
violet germicidal inactivation (UVGI). Due to the novelty ofthis
coronavirus strain at the time of writing, there are significant
gaps in literature on the UV susceptibility of the pathogen. In
this paper, estimates of the SARS-CoV-2 UVGI response are
derived from studies reporting UV susceptibility of SARS-CoV-
1, a close genomic relative. To motivate this comparison, the
genome sequences of both coronavirus strains were analyzed and
found to have effectively identical theoretical UV-C susceptibility,
differing by only 1.48%. Conducting a curve fitting analysis on
SARS-CoV-1 survivorship data obtained from existing literature,
the approximate UV-C dosage required to inactivate the virus
below assay limit of detection was found to be 36144 J/m2(≥5-
log). Using this dosage as a benchmark for UVGI applications
against SARS-CoV-2, a benchmark minimum exposure time can
be determined by t≈(1.5×106)π·(r2/P),whereris the
distance from the UV-C source to the sample surface, Pis the
wattage of the germicidal bulb, and tis expressed in seconds. This
amounts to at least 2 hours for a 15 W UV-C bulb placed 6 inches
from a disinfecting surface. The intent of this paper is to provide
readers with a tool to evaluate the effectiveness of a simple
UVGI system without the need for complex instrumentation in
an effort to guide COVID-19 pandemic mitigation efforts and
inform operational policy.
Index Terms—COVID-19, SARS-CoV-2, UV-C, dosage, expo-
sure time, UVGI
I. INTRODUCTION AND BACKGROUND
SARS-CoV-2, the highly contagious virus that caused the
COVID-19 global pandemic, is believed to be spread primarily
through droplet and contact transmission [1]. The former oc-
curs through direct or close contact with an infected individual.
The latter occurs when an infected individual leaves viral
particles on a surface or fomite. When another person comes
into contact with the same surface, remnants of the viable
virus may enter their respiratory tract, causing the second
person to become infected. As the demand for home delivery
spikes due to social distancing measures during the pandemic,
the transmission of viable viral loads through cardboard and
plastic packaging remains a significant concern. A recent study
[2] found that remnants of SARS-CoV-2 were found to live
on cardboard for over 24 hours and plastic for 3 days.
The potential for contact transmission has sparked
widespread interest in effective and reliable sterilization meth-
ods, including ultraviolet germicidal inactivation (UVGI), a
disinfection method that uses ultraviolet light in the 280 nm
-100nmrange(UV-C)toinactivatepathogens.Excitations
at these wavelengths can be absorbed by genetic material,
introducing mutations that can ultimately inactivate the or-
ganism [3]. While UVGI has conventionally been used in
air purification and water treatment systems [3], the current
pandemic has spurred a number of novel applications for this
technology such as the mass disinfection of public buses [4]
and airplane cabins [5], autonomous robots that enter poten-
tially infected hospital rooms [6], and chambers to sterilize
N95 masks for reuse in healthcare settings in order to manage
personal protective equipment (PPE) scarcity [7].
Due to the novelty of this coronavirus strain at the time of
writing, there have been no studies reporting the inactivation
effects of UV-C radiation on SARS-CoV-2. However, there is
abodyofliteraturereportinginactivationdosimetryforSARS-
CoV-1, the coronavirus strain responsible for the 2002 SARS
outbreak. To motivate the arguments and comparisons made
in this paper, a genomic analysis is conducted (Section II) to
establish that SARS-CoV-1 is a viable proxy for the novel
coronavirus in the context of UV-C inactivation. Applying
curve fitting techniques to existing literature on SARS-CoV-1,
an equation is proposed to determine the amount of exposure
time required to effectively inactivate the novel coronavirus
given UV-C bulb wattage and distance (Section III). The intent
of this paper is to make clear the extent of the effectiveness of
UVGI on the novel coronavirus in an effort to guide COVID-
19 pandemic mitigation efforts and inform operational policy.
II. GENOMIC COMPARISON OF SARS-COVSTRAINS
When absorbed by genetic material, excitation at UV-C
wavelengths can cause cross-linking in adjacent pyrimidine
bases. This phenomenon, called pyrimidine dimerization, can
inhibit successful replication and eventually lead to inactiva-
tion [3], [8]. Since these mutations have a higher probability of
occurring in regions of the genetic sequence with consecutive
pyrimidine bases, a dimerization probability can be derived
by counting such instances in the genomic sequence. This
metric can then be used to quantify UV-C susceptibility in
viruses. Kowalski [3], [8] gives the following equation to
predict dimerization probability Dvfor single-stranded RNA
(ssRNA) viruses such as SARS-CoV-2:
Dv=!"tt +α"←→
ct +β"cc +γ"←− −→
YYU
3
!n2
bp
(1)
where "tt,"cc,and"←→
ct represent all instances of ad-
jacent pyrimidines (i.e., TT,CC,andCT/TC,respectively)in
the genomic sequence; "←− −→
YYU represents instances where a
2
pyrimidine doublet is adjacent to an adenine or guanine purine
(e.g., ATT and GTC); nbp is the number of base pairs; and α,
β,andγare dimer proportionality constants. Suggested values
are α=0.1,β=6,andγ=4based on a linear fit analysis
of 28 ssRNA viruses [3], [8].
The genome sequences of both viruses were collected
from the National Institutes of Health (NIH) genetic se-
quence database (NC 004718 for SARS-CoV-1, NC 045512
for SARS-CoV-2) [9] and analyzed in a Python script accord-
ing to (1) using the suggested ssRNA dimer proportionality
constants. The results of this analysis are summarized in Table
I.
TAB L E I
GENOMIC COMPARISON OF SARS-COVSTRAINS
SARS-CoV-1 SARS-CoV-2
!tt 2207 2454
!cc 850 784
!←→
ct 3744 3494
!←− −→
YYU 7732 7736
nbp 29751 29903
Dv0.2458 0.2422
The two pathogens are close genomic relatives with 79%
sequence identity [10], suggesting that the strains have compa-
rable UV-C inactivation responses. The calculated dimerization
probabilities of the two coronavirus strains were found to be
effectively identical, differing by only 1.48%. These results
suggest that SARS-CoV-1 is a sound proxy for SARS-CoV-2
in the context of UV-C inactivation.
III. UV-C INACTIVAT ION OF SARS-COV-2
For a given UV-C dosage D(expressed in J/m2)viral
concentration decays exponentially as a function of time.
Useful parameters such as applied dosage can be numerically
extracted from experimental data by fitting the survivorship
curve to a decaying exponential with the form:
S=e−KIt (2)
where Sis the surviving fraction of the original virus popula-
tion after UV-C radiation is applied, Kis a factor called UV
susceptibility, and Iis the intensity at the sample surface [11].
UV-C dosage Dis related to susceptibility Kby:
D=−ln S
K(3)
Combining (2) and (3) yields the simplified relation:
D=It (4)
This numerical method is applied to multiple studies to
extract the approximate dosages for various stages of inacti-
vation. The results are tabulated in Table II. Each study either
directly reports a dosage or the measured intensity of the UV-
Csourceattheviralsample,inwhichcase(4)isusedto
calculate approximate dosage. Values in the first three rows
were taken from a literature review by Kowalski [17] which
tabulates experimentally derived UV-C dosages across various
TAB L E I I
LITERATURE REVIEW OF SARS-COV-1 I NACTIVAT ION DOSAGES
Study D90 D99.9D99.99 D99.999
Wal ke r 20 07 [1 2] 7 - - -
Duan 2003 [13] 9 - - -
Kariwa 2006 [14] - 134 - -
Eickmann 2020 [15] - 5000 10000-15000 -
Darnell 2004 [16] 2410 6020 12050 36144
*valuesareapproximateandexpressedinJ/m
2
SARS-CoV-1 papers. The D99 column was omitted due to
insufficient data across all five studies. The largest reported
dosage D= 36144 J/m2will be used here as the target UV-C
inactivation dosage against SARS-CoV-2.
IV. DETERMINING MINIMUM EXPOSURE TIME
In this section, a simple approximation to determine min-
imum exposure time is derived and presented. In an effort
to help guide operational policy, conservative assumptionsare
made where necessary in order to express the equation in terms
of known parameters such as the wattage of the bulb used and
distance to the sample. Minimum exposure time is determined
by solving (4) for t,givenD= 36144 J/m2.Intensityatthe
sample surface Iis approximated by:
I=PUVC
Aexposed
=ηP
4πr2(5)
where Pis the power rating of the germicidal bulb (typically
either 15 W or 36 W), ηis an attenuation factor where η≤1,
and ris the distance from the bulb to the sample.
For a conventional 15 W germicidal bulb, only about one-
third (5 W) of the original power Pis dissipated as UV-C
radiation [18]. Additionally, depending on the placement and
shape of the bulb, only a fraction of the emitted UV-C radiation
will hit the surface. Forgoing further analysis, a conservative
and reasonable estimate used here is η≈10%.Theexposed
area Aexposed is estimated using spherical spreading, where r
is the distance from the UV-C source to the irradiated surface.
Combining (4) and (5) yields an equation to determine
approximate exposure time tto achieve the target inactivation
dosage D= 36144 J/m2for SARS-CoV-2:
t≈1.5×106#πr2
P$(6)
where Pis the power rating of the germicidal bulb, ηis a
coefficient representing heat and scattering losses, and tis
expressed in seconds. Using (6), a simple UVGI chamber setup
consisting of a 15 W UV-C bulb placed 0.15 meters (6 inches)
away from the sanitizing surface will require at least 2 hours
of exposure to achieve the target inactivation dosage.
V. D ISCUSSION
The Centers for Disease Control and Prevention (CDC)
recommend a minimum UV-C dosage of 10000 J/m2[19] for
N95 mask sterilization based on studies that report significant
viral inactivation (≥3-log) at this dosage. Several of these
studies were omitted from the literature review here because
3
the sample pathogen was not SARS-CoV-1. Nevertheless,
the results of the literature review conducted here are in
close agreement with those studies and reinforces the CDC
recommendation.
However, the expression of UVGI effectiveness in terms
of dosage may be too much of an abstraction for some
readers. The approximation in (6) provides readers with a
tool to evaluate the effectiveness of a simple UVGI system
based on bulb wattage and distance to the sample surface
without the need for complex instrumentation. The target
dosage D= 36144 J/m2derived here has been shown to
inactivate 99.999% (≥5-log) of SARS-CoV-1 contaminants in
alaboratorysettingandprovidessignificantmarginrelative to
the minimum CDC recommendation of D= 10000 J/m2.
A. Limitations of UVGI
As a number of studies have noted, there are many factors
that can impede UVGI ability in practice [20], [21]. For
N95 respirator disinfection applications, there are significant
concerns as to whether UVGI systems can deliver the sufficient
dosage to inactivate viral particles trapped within the fibers of
the mask surface [22]. Shadowing effects must also be taken
into consideration when designing UVGI systems [23].
It is important to note that while [16] demonstrated a 5-
log reduction in SARS-CoV-1 concentration, this should not
be taken to mean that sterilized surfaces no longer pose an
infectivity threat. For surfaces containing extremely highviral
loads, for instance, 99.999% inactivation does not necessar-
ily translate to complete sterilization. In one recent study,
researchers reported SARS-CoV-2 viral loads (expressed in
copies/mL) from COVID-19 patient samples ranging from
641 to 1.34 ×1011 with a reported median of 7.52 ×105
[24]. Supposing a surface is infected with the median value,
atheoretical5-logreductioninvirusconcentrationwould
decrease the viral load to 100 copies/mL. However, taking the
upper bound of the dataset would still leave a viable residual
viral load of 106copies/mL.
While UVGI has been shown to reduce the risk of contact
transmission, more robust disinfection protocols can be de-
signed by implementing a multimodal approach that provides
redundancy if environmental factors do not allow for complete
UV-C inactivation. The CDC recommends the use of UVGI
methods in conjunction with vaporous hydrogen peroxide
treatment and moist heat [19]. A well designed protocol that
uses UVGI in conjunction with other proven disinfection
methods can be a good way to manage and mitigate viral
infectivity risk during the COVID-19 pandemic.
REFERENCES
[1] “Modes of transmission of virus causing COVID-19: Implications for
IPC precaution recommendations: scientific brief, 27 March 2020,”
Wor ld H eal th O rg ani za tio n, Tech . Re p., 2 020 .
[2] N. van Doremalen, T. Bushmaker et al.,“Aerosolandsurfacestability of
SARS-CoV-2 as compared with SARS-CoV-1,” New England Journal
of Medicine,2020.
[3] W. Kowalski, Ultraviolet germicidal irradiation handbook: UVGI for air
and surface disinfection.SpringerScience&BusinessMedia,2010.
[4] Z. Gorvett, “Can you kill coronavirus with UV light?” BBC News,April
2020.
[5] S. K. Moore, “Flight of the GermFalcon: How a Potential Coronavirus-
Killing Airplane Sterilizer Was Born,” IEEE Spectrum: Technology,
Engineering, and Science News,March2020.
[6] E. Ackerman, “Autonomous Robots Are Helping Kill Coronavirus in
Hospitals,” IEEE Spectrum: Technology, Engineering, and Science News,
March 2020.
[7] K. J. Card, D. Crozier et al.,“UVSterilizationofPersonalProtective
Equipment with Idle Laboratory Biosafety Cabinets During the COVID-
19 Pandemic,” medRxiv,2020.
[8] W. J. Kowalski, W. P. Bahnfleth, and M. T. Hernandez, “A Genomic
Model for the Prediction of Ultraviolet Inactivation Rate Constants for
RNA and DNA Viruses,” IUVA News,June2009.
[9] K. Clark, I. Karsch-Mizrachi, D. J. Lipman, J. Ostell, andE.W.Sayers,
“GenBank,” Nucleic acids research,vol.44,no.D1,pp.D67–D72,2016.
[10] P. Zhou, X.-L. Yang et al.,“Apneumoniaoutbreakassociatedwitha
new coronavirus of probable bat origin,” Nature,vol.579,no.7798,pp.
270–273, 2020.
[11] C. Tseng and C. Li, “Inactivation of viruses on surfaces by ultraviolet
germicidal irradiation,” Journal of Occu pationa l an d Envi ronmenta l
Hygiene,vol.4,no.6,pp.400–405,2007.
[12] C. M. Walker and G. Ko, “Effect of ultraviolet germicidalirradiationon
viral aerosols,” Environmental Science & Technology,vol.41,no.15,
pp. 5460–5465, 2007.
[13] S. Duan, X. Zhao, R. Wen, J.-j. Huang, G. Pi, S. Zhang, J. Han,
S. Bi, L. Ruan, and X.-p. Dong, “Stability of SARS coronavirusin
human specimens and environment and its sensitivity to heating and
UV irradiation,” Biomedical and Environmental Sciences: BES,vol.16,
no. 3, pp. 246–255, 2003.
[14] H. Kariwa, N. Fujii, and I. Takashima, “Inactivation of SARS coron-
avirus by m eans of povidone-iodine, phys ical conditions andchemical
reagents,” Dermatology,vol.212,no.Suppl.1,pp.119–123,2006.
[15] M. Eickmann, U. Gravemann, W. Handke, F. Tolksdorf, S. Reichen-
berg, T. H. M¨uller, and A. Seltsam, “Inactivation of three emerging
viruses–severe acute respiratory syndrome coronavirus, Crimean–Congo
haemorrhagic fever virus and Nipah virus–in platelet concentrates by
ultraviolet C light and in plasma by methylene blue plus visible light,”
Vox Sa n g u inis .
[16] M. E. Darnell, K. Subbarao, S. M. Feinstone, and D. R. Taylor,
“Inactivation of the coronavirus that induces severe acute respiratory
syndrome, SARS-CoV,” Journal of Virological M ethods,vol.121,no.1,
pp. 85–91, 2004.
[17] W. Kowalski, T. Walsh, and V. Petraitis, “2020 COVID-19 Coronavirus
Ultraviolet Susceptibility,” March 2020.
[18] Germicidal Low Pressure Mercury Arc,UshioAmerica,Inc.,5440
Cerritos Avenue, Cypress, CA, 2019.
[19] “Recommended Guidance for Extended Use and Limited Reuse of N95
Filtering Facepiece Respirators in Healthcare Settings,” April 2020.
[20] G. Byrns, B. Barham et al.,“Theusesandlimitationsofahand-
held germicidal ultraviolet wand for surface disinfection,” Journal of
Occupational and Environmental Hygiene,vol.14,no.10,pp.749–757,
2017.
[21] M. Lindblad, E. Tano, C. Lindahl, and F. Huss, “Ultraviolet-C decon-
tamination of a hospital room: Amount of UV light needed,” Burns,
2019.
[22] E. Fisher and R. Shaffer, “A method to determine the available UV-C
dose for the decontamination of filtering facepiece respirators,” Journal
of Applied Microbiology,vol.110,no.1,pp.287–295,2011.
[23] D. Mills, D. A. Harnish, C. Lawrence, M. Sandoval-Powers, and B. K.
Heimbuch, “Ultraviolet germicidal irradiation of influenza-contaminated
N95 filtering facepiece respirators,” American journal of infection con-
trol,vol.46,no.7,pp.e49–e55,2018.
[24] Y. Pan, D. Zhang, P. Yang, L. L. Poon, and Q. Wang, “Viral load
of SARS-CoV-2 in clinical samples,” The Lancet Infectious Diseases,
vol. 20, no. 4, pp. 411–412, 2020.
