Water Research 94 (2016) 341e349

Contents lists available at ScienceDirect

Water Research
journal homepage: www.elsevier.com/locate/watres

Review

Application of ultraviolet light-emitting diodes (UV-LEDs) for water

disinfection: A review
Kai Song, Madjid Mohseni, Fariborz Taghipour*
Department of Chemical and Biological Engineering, The University of British Columbia, 2360 East Mall, Vancouver, BC V6T 1Z3, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Ultraviolet (UV) disinfection is an effective technology for the inactivation of pathogens in water and is of
Received 28 October 2015 growing interest for industrial application. A new UV source d ultraviolet light-emitting diode (UV-LED)
Received in revised form d has emerged in the past decade with a number of advantages compared to traditional UV mercury
29 February 2016
lamps. This promising alternative raises great interest in the research on application of UV-LEDs for
Accepted 1 March 2016
Available online 2 March 2016
water treatment. Studies on UV-LED water disinfection have increased during the past few years. This
article presents a comprehensive review of recent studies on UV-LEDs with various wavelengths for the
inactivation of different microorganisms. Many inconsistent and incomparable data were found from
Keywords:
UV disinfection
published studies, which underscores the importance of establishing a standard protocol for studying
Ultraviolet light-emitting diode (UV-LED) UV-LED inactivation of microorganisms. Different UV sensitivities to UV-LEDs and traditional UV lamps
Water treatment were observed in the literature for some microorganisms, which requires further investigation for a
Inactivation effectiveness better understanding of microorganism response to UV-LEDs. The unique aspects of UV-LEDs improve
inactivation effectiveness by applying LED special features, such as multiple wavelengths and pulsed
illumination; however, more studies are needed to investigate the influencing factors and mechanisms.
The special features of UV-LEDs offer the flexibility of novel reactor designs for a broad application of UV-
LED reactors.
© 2016 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
2. Inactivation effectiveness of UV-LEDs at various wavelengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
3. Microorganism response to UV-LEDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
4. Time-response data on UV-LED disinfection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
5. Mechanism of inactivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
6. Future research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348

1. Introduction people around the world lack access to a safe drinking water source
and are threatened by waterborne diseases annually (Hatami, 2013;
Drinking water safety is an important issue worldwide, espe- WHO, 2014). The development of efficient water treatment tech-
cially in most developing countries and rural areas. Millions of nologies, especially inactivation of pathogenic microorganisms in
water, is of great significance for human health and well-being.
Ultraviolet (UV) radiation can effectively inactivate various mi-
croorganisms in water (Hijnen et al., 2006) and has been increas-
* Corresponding author.
ingly used for water disinfection. UV radiation has numerous
E-mail address: fariborz.taghipour@ubc.ca (F. Taghipour).

http://dx.doi.org/10.1016/j.watres.2016.03.003
0043-1354/© 2016 Elsevier Ltd. All rights reserved.
342 K. Song et al. / Water Research 94 (2016) 341e349
advantages over conventional chemical disinfection (e.g., chlori-
nation or ozonation), such as no chemical addition, no harmful
disinfection by-products (DBPs) formation, and no introduction of
disinfectant-resistance to bacteria (Mori et al., 2007). UV disinfec-
tion has been recommended as a substitute for chemical additives
for surface water treatment (USEPA, 2006). Currently, there are
more than 7000 municipal UV disinfection installations in the
world (Muller, 2011), and small household UV disinfection systems
are also available (Brownell et al., 2008). It is estimated that the
global market for UV disinfection equipment has a potential to
reach $2.8 billion by 2020 (Allied Analytics LLP, 2014). The main UV
sources for current UV disinfection systems are low- or medium-
pressure mercury lamps (Chevremont et al., 2013a). Although
these lamps are widely used in water treatment systems, there are
still many issues with them. The major concern is that the UV lamps
are fragile and contain toxic mercury, which is hazardous to the
environment and requires proper disposal (Chevremont et al.,
2013b; Close et al., 2006). Moreover, these lamps require signifi-
cant amounts of energy to operate due to a low wall plug efficiency
of around 15e35% and have a relatively short lifetime of about
10,000 h (Autin et al., 2013; Chatterley and Linden, 2010).
In the past few years, with the rapid development and
improvement of the semiconductor industry, UV light-emitting
diodes (UV-LEDs) have emerged as a new source for UV radiation
generation. An LED is a semiconductor device that utilizes semi-
conducting materials to create a pen junction (hole and electron).
The electrons and holes recombine at the junction to emit radia-
tion, and the wavelength of the radiation depends on the semi-
conducting materials. Commercial visible LEDs have been available
for nearly 50 years and have diverse applications, especially in the
lighting industry, due to the increasingly higher efficiency and
lower cost (Ibrahim et al., 2014). Recently, the newly emerging UV-
LEDs have followed a similar track and are expected to be
economically viable in the coming years (Harris et al., 2013).
UV-LEDs at various wavelengths can be manufactured using
different semiconducting materials. The most frequently used
materials are III-nitride, including gallium nitride (GaN),
aluminium gallium nitride (AlGaN), and aluminum nitride (AlN)
(Khan et al., 2005). The wavelength of GaN-based UV-LEDs can be
as short as 365 nm, which is in the near UV region (Taniyasu et al.,
2006b). However, the AIN UV-LEDs are reported to emit UV radi-
ation at 210 nm (deep UV), which is the shortest wavelength among
semiconductors (Taniyasu et al., 2006a). A wavelength from 210 to
365 nm (covering from deep UV to near UV regions) is available
from the emission of AlGaN, which consists of AIN and GaN in
appropriate proportions (Taniyasu and Kasu, 2010). Because the
wavelength is found to be an essential factor for water disinfection
efficiency (Vilhunen et al., 2009), the ability of UV-LEDs to offer a
variety of wavelengths is well aligned with the needs of efficient
disinfection, making it a potential option.
In addition to diversity in wavelengths, UV-LEDs possess several
unique advantages such as environmental friendliness (no mer-
cury), compactness and robustness (more durable), faster start-up
time (no warm-up time), potentially less energy consumption,
longer lifetime, and the ability to turn on and off with high fre-
quency (Wurtele et al., 2011). It is predicted that by 2020, UV-LEDs
will operate at 75% wall plug efficiency with a lifetime longer than
100,000 h, comparable to the operating parameters of current
visible LEDs (Autin et al., 2013; Ibrahim et al., 2014). All these fac-
tors make UV-LEDs a promising alternative to conventional UV
mercury lamps for water disinfection. However, due to the sub-
stantial differences between the traditional mercury lamps and
newly emerging UV-LEDs, the numerous research results and
established methodologies on water disinfection by mercury
lamps, such as experimental protocols, reactor designs, and
inactivation kinetics, may not apply directly to UV-LEDs. Moreover,
few studies currently exist that investigate the application of UV-
LEDs to water disinfection. Therefore, a comprehensive research
on UV-LEDs for water disinfection is essential to better understand
the feasibility and future applications of this technology.
This review presents and discusses published data from the
literature, aiming to give an overall understanding of UV-LED water
disinfection. Using the results of this review, some of the main
future research directions are identified. To the best of the authors'
knowledge, this is the first review paper on the application of newly
emerging UV-LEDs for water disinfection.
2. Inactivation effectiveness of UV-LEDs at various
wavelengths
There has been a lack of uniformity in research materials and
methods reported for UV-LED disinfection studies over the last
decade, making comparisons difficult. Table 1 summarizes the main
results from the published work on UV-LED water disinfection. The
data on UV dose and log inactivation were obtained from each
study, and the UV dose response, as the value of UV dose/log inac-
tivation (unit: mJ/cm2 per log inactivation), was calculated to
evaluate and compare the effectiveness of disinfection among
different studies.
The germicidal efficiency of UV radiation was reported to be
highly dependent on the wavelength, and the spectral sensitivity of
microorganisms was found to not necessarily follow the DNA
absorbance spectrum (Chen et al., 2009; Mamane-Gravetz et al.,
2005). Therefore, UV wavelength is an essential factor for micro-
organism inactivation, and the effectiveness may vary from
microorganism to microorganism (Linden et al., 2001; Vilhunen
et al., 2009). Different UV wavelengths in the range of
315e400 nm (UVA), 280e315 nm (UVB), and <280 nm (UVC) have
been applied for microorganism inactivation by UV-LEDs. From the
published data, sorted by wavelength (Table 1), the influence of UV-
LED wavelengths on inactivation effectiveness can be evaluated.
Hamamoto et al. (2007) and Mori et al. (2007) applied UVA
radiation with 365-nm UVA-LEDs on Escherichia coli and obtained
UV dose responses of 55,263 and 13,846 mJ/cm2 per log inactiva-
tion, respectively (Table 1). These values are high considering that
typically the UV dose required by 254-nm mercury lamps for 4 log
E. coli inactivation is about 8 mJ/cm2, which makes the UV dose
response only 2 mJ/cm2 per log inactivation (Bolton and Cotton,
2008). With the help of titanium dioxide (TiO2), Xiong and Hu
(2013) established a photocatalytic disinfection system using 365-
nm UV-LEDs, and the results still showed a high UV dose
response of 229 mJ/cm2 per log inactivation for E. coli inactivation.
These results demonstrated that 365-nm UVA-LEDs alone are not
effective for microorganism inactivation.
Oguma et al. (2013) used 310-nm UVB-LEDs for E. coli inacti-
vation and reported a UV dose response of 94.8 mJ/cm2 per log
inactivation, which is much lower than that required for 365-nm
UVA-LEDs. Nonetheless, this result was far greater than the
values reported for UVC-LEDs for various microorganisms, which
ranged from 1.0 to 30.5 mJ/cm2 per log inactivation (Table 1).
Therefore, UVB-LEDs alone are also not highly effective for micro-
organism inactivation.
Aoyagi et al. (2011) selected 255-nm and 280-nm UVC-LEDs to
study the inactivation effects on bacteriophages 4X174, Qb, and
MS2. The results showed that the UV dose responses at 280 nm for
these three microorganisms were 2.8, 28.7, and 30.5 mJ/cm2 per log
inactivation, respectively (Table 1), which were all higher than
those at 255 nm (1.7, 12.5, and 12.8 mJ/cm2 per log inactivation,
respectively). The results indicated that UVC-LEDs at 255 nm could
be more effective than at 280 nm for the inactivation of
 K. Song et al. / Water Research 94 (2016) 341e349 343

Table 1
Summary of dose response data for water disinfection by UV-LEDs.

Wavelength Microorganism Disinfection UV dose (mJ/ Log Dose response (mJ/cm2 per log Reference
(nm) medium cm2) inactivation inactivation)

250 B. subtilis Water 59.2 3 19.7 Morris, 2012

254 Mesophilic bacteria Water 0.73 0.8 1.0 Chevremont et al., 2012a
255 4X174 Water 6.4 3.7 1.7 Aoyagi et al., 2011
255 Qb Water 30 2.4 12.5 Aoyagi et al., 2011
255 MS2 Water 41 3.2 12.8 Aoyagi et al., 2011
255 MS2 Water 60 2.3 26.1 Bowker et al., 2011
255 T7 Water 20 3.9 5.1 Bowker et al., 2011
255 E. coli Water 9 2.7 3.3 Bowker et al., 2011
265 E. coli Water 20 3.4 5.9 Chatterley and Linden,
2010
265 E. coli Water 10.8 4 2.7 Oguma et al., 2013
265 Pseudomonas Biofilm in tube 7.8 4 2.0 Bak et al., 2010
aeruginosa
269 B. subtilis Water 40 5.9 6.8 Wurtele et al., 2011
275 MS2 Water 60 2.1 28.6 Bowker et al., 2011
275 T7 Water 20 4.7 4.3 Bowker et al., 2011
275 E. coli Water 9 3.8 2.4 Bowker et al., 2011
280 E. coli Water 13.8 4 3.5 Oguma et al., 2013
280 Mesophilic bacteria Water 1.37 1.4 1.0 Chevremont et al., 2012a
280 4X174 Water 8.9 3.2 2.8 Aoyagi et al., 2011
280 Qb Water 43 1.5 28.7 Aoyagi et al., 2011
280 MS2 Water 58 1.9 30.5 Aoyagi et al., 2011
282 B. subtilis Water 60 7.2 8.3 Wurtele et al., 2011
310 E. coli Water 56.9 0.6 94.8 Oguma et al., 2013
365 E. coli Water 315,000 5.7 55,263 Hamamoto et al., 2007
365 E. coli Water 54,000 3.9 13,846 Mori et al., 2007
365 E. coli Water/TiO2 688 3 229 Xiong and Hu, 2013
365 Mesophilic bacteria Water 4.22 0.3 12.5 Chevremont et al., 2012a
405 Mesophilic bacteria Water 25.58 0.3 88.0 Chevremont et al., 2012a
254/365 Mesophilic bacteria Water 4.95 2.4 2.1 Chevremont et al., 2012a
280/365 Mesophilic bacteria Water 5.59 3.5 1.6 Chevremont et al., 2012a
254/405 Mesophilic bacteria Water 26.31 2.2 11.9 Chevremont et al., 2012a
280/405 Mesophilic bacteria Water 26.95 3.5 7.7 Chevremont et al., 2012a

bacteriophages 4X174, Qb, and MS2.
Another study on UVC-LEDs at 255 nm and 275 nm was re-
ported by Bowker et al. (2011) for the inactivation of three micro-
organisms: MS2, T7, and E. coli. Similar UV dose responses were
obtained from this study for the 255-nm and 275-nm wavelengths
(26.1, 5.1, and 3.3 mJ/cm2 per log inactivation versus 28.6, 4.3, and
2.4 mJ/cm2 per log inactivation for MS2, T7, and E. coli, respectively;
Table 1). For MS2, the UV dose response at 255 nm was slightly
lower than that at 275 nm, indicating that UVC-LED at 255 nm was
more effective than at 275 nm. However, for T7 and E. coli, 255 nm
resulted in slightly higher UV dose responses than at 275 nm,
suggesting that 275 nm was more effective. The results for MS2 and
T7 were consistent with their action spectra, considering that the
action spectrum of MS2 has a peak around 260 nm and the peak for
T7 is around 270 nm (Mamane-Gravetz et al., 2005; Ronto et al.,
1992; Beck et al., 2015). The results did not agree with the E. coli
action spectrum given that 255 nm is closer to the peak around
260e265 nm and therefore 255 nm is expected to be more effective
than 275 nm (Bolton and Cotton, 2008). The higher effectiveness of
275 nm may be attributed to the higher output power of the 275-
nm UV-LEDs, resulting in a higher fluence rate and shorter expo-
sure time to reach the same UV dose, which is proposed to be more
effective for E. coli inactivation than the combination of a lower
fluence rate and longer exposure time (Bowker et al., 2011; Harm,
1980; Sommer et al., 1998). Basically, UV inactivation of microor-
ganisms is believed to follow the BunseneRoscoe reciprocity law,
which states that the photochemical effect depends only on the
total energy dose, i.e., the product of fluence rate and exposure time
(Murata and Osakabe, 2013). This was verified by Sommer et al.
(1998) who reported that for inactivation of bacteriophages MS2,
4X174, and B40-8 by mercury UV lamps, log inactivation is the
same as long as the product of fluence rate and exposure time is the
same. At the same time, these authors found a deviation from the
reciprocity law for E. coli inactivation that showed a higher fluence
rate and shorter exposure time resulted in a higher log inactivation
than a lower fluence rate and a longer exposure time despite the
total UV fluence (UV dose) being the same. This phenomenon could
be attributed to UV disinfection depending not only on photo-
chemical reactions but also on biological processes (Harm, 1980).
Therefore, the deviation from the reciprocity law may occur on
certain microorganisms, like E. coli, because the biological pro-
cesses induced by UV radiation may vary with different microor-
ganisms, and some of the organisms may be more sensitive to the
fluence rate.
Although some of the UV-LED dose responses reported from
different studies are in agreement, there are many cases where the
results are inconsistent. One such example is the inconsistency in
results between studies by Aoyagi et al. (2011) and Bowker et al.
(2011). Both studies used 255-nm UVC-LEDs for MS2 inactivation.
However, in the former case, a 41-mJ/cm2 UV dose provided a 3.2
log inactivation, whereas in the latter, a 60-mJ/cm2 UV dose ach-
ieved only 2.3 log inactivation. Similar inconsistent results were
also observed between the studies by Chatterley and Linden (2010)
and Oguma et al. (2013), where both applied 265 nm to E. coli and
reported UV dose responses of 5.9 and 2.7 mJ/cm2 per log inacti-
vation, respectively. These disagreements might result from the
differences among materials and experimental conditions in the
different studies. Different UV-LEDs in these studies have various
radiation patterns, such as emission spectra, viewing angle, and
radiation distribution. Various apparatuses were applied for UV-
LED water disinfection, and different methods were used to
determine the UV dose from the UV-LEDs. Currently, there is no
344 K. Song et al. / Water Research 94 (2016) 341e349
(A)
(B)
UV lamp
Collimated UV-LED
beam
Water sample
container and
Water sample stirrer
container and
stirrer
Fig. 1. Schematic diagram for typical UV lamp collimated beam apparatus (A) and UV-
LED bench scale apparatus (B).
consistent methodology for obtaining the UV-LED dose response of
microorganisms, and there is no standard protocol for determining
the UV dose delivered by UV-LEDs to a sample solution. Therefore,
the discrepancy among different studies is inevitable.
The standardized protocol for microorganism inactivation by
conventional UV mercury lamps has been well established, and the
results under the standardized procedure are comparable (Bolton
and Linden, 2003; Kuo et al., 2003). However, the UV lamp proto-
col is not expected to be directly applicable to UV-LEDs because of
the substantial differences between mercury lamps and UV-LEDs.
The bench-scale UV disinfection experiments for mercury lamps
usually utilize a collimated beam apparatus (Fig. 1A). The key part
of this apparatusda collimated beamdis used to provide a uniform
irradiation field on a surface area by collimating a parallel beam to
vertically project on the water surface, so that the irradiance on the
surface of the sample can be accurately measured by a radiometer,
enabling the UV dose to be determined with necessary corrections
(Bolton and Linden, 2003). To ensure the uniformity of the beam
and accuracy of the measurement, the length of the collimated
beam has to be at least 20 cm (Blatchley, 1997). Such a long colli-
mated beam is not practical for UV-LEDs due to their low radiant
power. Currently, the output power of a UVC-LED is only several
milliwatts, which is much lower than that of low-pressure mercury
UV lamps (typically 40 W) or high-output medium-pressure mer-
cury lamps (up to 30 kW). As a result, the UV-LEDs have to be close
to a water sample in which inactivation is performed for deliverable
UV energy. Various apparatus can be applied for UV-LED inactiva-
tion of microorganisms. Typically, the UV-LEDs are located directly
above the water sample (Fig. 1B), and the distance between UV-
LEDs and the water sample is usually no more than 2 cm
(Chevremont et al., 2012a, 2012b; Hamamoto et al., 2007; Hwang
et al., 2013; Li et al., 2010; Mori et al., 2007; Oguma et al., 2013;
Vilhunen et al., 2009). Considering that a UV-LED is a point
source and emits hemispheric radiation, the UV emission from a
UV-LED is neither parallel nor vertical to the water surface within
such a short distance between the UV-LED and the water sample
(Fig. 1B). As a result, uniform irradiance is not expected on the
water sample surface, which leads to difficulties for the accurate
determination of UV dose. Further, the radiant power of UV-LEDs
can be significantly affected by operating parameters, such as the
applied current and voltage, and by thermal management during
the operation. Therefore, there is a need for the development of a
standardized protocol for a UV-LED-based inactivation study of
microorganisms, especially for the accurate determination of UV
dose and the proper operation of the system. This necessity has
been recognized by researchers and industry leaders in the field of
UV-LEDs. Recently, an International Ultraviolet Association (IUVA)
initiative has been announced, undertaken by a working group of
the IUVA Manufacturers Council, to present a consistent method-
ology for the determination and benchmarking of UVC output from
LEDs (IUVA, 2015).
Chevremont et al. (2012a) studied the inactivation effect of
coupled UVA- and UVC-LEDs on mesophilic bacteria. The UV dose
responses for UVA alone at 365 nm and 405 nm were 12.5 and
88 mJ/cm2 per log inactivation, which are much higher than those
for the UVC alone at 254 nm and 280 nm (both 1.0 mJ/cm2 per log
inactivation). However, when combining UVA- and UVC-LEDs, the
UV dose responses at 254/365 nm, 280/365 nm, 254/405 nm, and
280/405 nm all sharply decreased (2.1, 1.6, 11.9, and 7.7 mJ/cm2 per
log inactivation, respectively), indicating that the combination of
UVA- and UVC-LEDs is more efficient than UVA-LEDs alone.
Moreover, in terms of log inactivation, it was found that the com-
bination of 280 nm and 365 nm provided higher log inactivation
than the sum of each alone (3.5 > 1.4 þ 0.3 ¼ 1.7). The other three
combinations showed similar phenomena: 254/365 nm
(2.4 > 0.8 þ 0.3 ¼ 1.1), 254/405 nm (2.2 > 0.8 þ 0.3 ¼ 1.1), and 280/
365 nm (3.5 > 1.4 þ 0.3 ¼ 1.7). This synergistic effect from the
wavelength combinations was also reported by Nakahashi et al.
(2014), who combined 254 nm with 365 nm for Vibrio para-
haemolyticus inactivation. However, Oguma et al. (2013) combined
265-nm, 280-nm, and 310-nm UV-LEDs for E. coli inactivation and
did not observe a synergistic effect. On the contrary, these re-
searchers found that combined wavelengths were less efficient
than each wavelength applied separately, which could have resul-
ted from different indicator microorganisms and wavelength
combinations, as well as an inefficient thermal management of the
experimental setup (Oguma et al., 2013). These studies suggest that
combination of selected wavelengths might be a promising way to
improve the disinfection efficiency of UV-LEDs, but more studies
are needed because the experimental setup and conditions may be
influential.
Although a medium-pressure mercury lamp can provide poly-
chromatic radiation, its spectrum, which includes UVA, UVB, UVC,
and visible light, is fixed and cannot be adjusted. Different effects
and mechanisms for medium-pressure mercury lamp inactivation
of microorganisms have been attributed to its polychromatic radi-
ation. However, within the very broad range of wavelengths in its
spectrum, it is difficult to identify and distinguish which wave-
length or which combination of specific wavelengths the additional
effects and mechanisms result from. UV-LEDs provide great flexi-
bility for wavelength combinations and fluence rate due to their
unique feature of wavelength diversity. This flexibility offers a new
approach to tailor wavelength combinations and the radiant power
of each peak wavelength for optimal inactivation and to identify the
additional mechanism of a particular combination and fluence rate.
The ability to turn on and off with a high frequency is another
unique feature of UV-LEDs, enabling adjustable UV pulsed illumi-
nation. Such a feature makes UV-LEDs desirable for potentially
enhancing the inactivation effectiveness by pulsed irradiation. Li
et al. (2010) reported enhanced germicidal effects of pulsed UV-
LED irradiation by applying UVA-LED at 365 nm for inactivation
of Candida albicans and E. coli biofilms. They found that pulsed
irradiation had significantly greater germicidal ability than
continuous irradiation with a maximum at 100 Hz and 75% duty
rate (the percentage of the exposure time in total operating time)
under the same UV dose. Wengraitis et al. (2013) applied pulsed
UVC radiation by UV-LEDs at 272 nm for E. coli disinfection on agar
plates and found it to be more effective than continuous illumi-
nation; the log inactivation for pulsed UVC radiation at 272 nm with
1 Hz and 10% duty rate was 3.8 times higher than that of continuous
illumination based on the same UV dose, indicating the high effi-
ciency of pulsed UV radiation.
 K. Song et al. / Water Research 94 (2016) 341e349 345

The conventional UV source for generating pulsed UV is a xenon inactivation.

lamp. The enhanced germicidal effect of xenon lamp-pulsed UV has
been demonstrated by studies for food decontamination and water
disinfection (Bohrerova et al., 2008; Elmnasser et al., 2007; Gomez-
Lopez et al., 2007; Oms-Oliu et al., 2010). However, the pulses
generated by a xenon lamp are different from those of UV-LEDs in
terms of spectrum, intensity, frequency, and duty rate. The wave-
length distribution of xenon lamp pulses ranges from 100 nm to
1000 nm, including UV, visible light, and infrared with a peak po-
wer up to 35 MW. Usually the duration of each pulse is from
nanoseconds to milliseconds, and typically 1 to 20 pulses are
emitted per second (Oms-Oliu et al., 2010). As discussed previously,
UV-LEDs have selectable wavelengths with relatively low radiant
power and are capable of highly adjustable pulsed illumination
with broad ranges for frequency and duty rate. The numerous
studies on xenon lamp-pulsed UV reveal that these pulse patterns
play important roles for enhanced germicidal effect (Elmnasser
et al., 2007). Therefore, due to the differences in pulse patterns,
the direct applicability of the findings on xenon lamp-pulsed UV to
UV-LED pulsed illumination is not expected, and more studies on
UV-LED pulsed illumination are needed to confirm the enhanced
germicidal effect.
3. Microorganism response to UV-LEDs
Although UV radiation is effective against most pathogenic
microorganisms in water, different microorganisms have different
responses to UV radiation due to various UV resistances and pro-
cess conditions. The sensitivity of microorganisms to UV radiation
can be evaluated by the inactivation rate constant k (cm2/mJ) from
the linear portion of the relationship between log inactivation and
the applied UV dose (Hijnen et al., 2006):
Log inactivation ¼ k  UV dose
The k values for various microorganisms were calculated based
on the data from each study (summarized in Table 1), and are listed
in Table 2. The data from conventional UV mercury lamps at 254 nm
are also shown in Table 2 for comparison. A high k value means the
microorganism is UV sensitive and requires a low UV dose for
The k values for bacteriophage 4X174 by 255-nm and 280-nm
UV-LEDs were reported as 0.578 and 0.360 cm2/mJ, respectively
(Aoyagi et al., 2011), which were higher than all other microor-
ganisms in Table 2, suggesting that 4X174 is very vulnerable to UVC
radiation. The k value for 4X174 by 254 nm was also much higher
than that for most of the other microorganisms in the studies
concerning UV lamps. However, the k value of UV lamps at 254 nm
was closer to that of UV-LEDs at 280 nm instead of 255 nm. This
disagreement may result from the differences between these two
UV sources, including the fact that the UV fluence rate of conven-
tional UV lamps is much higher than that for UV-LEDs, which may
affect the microorganism responses (Harm, 1980). On the other
hand, 4X174 inactivation by UV lamps showed no sensitivity to
different fluence rates based on the study by Sommer et al. (1998).
Nonetheless, further research is needed to investigate the UV-LED
wavelength effect given the dearth of literature on this aspect.
E. coli inactivation was investigated with UVC-LEDs at different
wavelengths (255, 265, 272, 275, and 280 nm) in several studies
(Bowker et al., 2011; Chatterley and Linden, 2010; Oguma et al.,
2013; Wengraitis et al., 2013). The k values ranged from 0.170 to
0.422 cm2/mJ, which were slightly lower than the values for 4X174,
indicating that E. coli is quite sensitive to UVC radiation. The wide
range of k values suggests that UV sensitivity of E. coli largely de-
pends on the wavelength. Moreover, although these studies used
E. coli as the indicator microorganism, the strains of E. coli were
different, which could be a factor that results in the different
inactivation rate constants because UV sensitivity may vary with
different strains of a species, especially for E. coli (Malley et al.,
2004; Sommer et al., 1998, 2000). The average k value from UV
lamp studies for E. coli inactivation was 0.506 cm2/mJ, which was
higher than the values found using UV-LEDs, with 275 nm resulting
in the closest value, rather than the 255-nm UV-LED. A possible
explanation may be the difference of fluence rate from UV lamps
and UV-LEDs since a few studies have found that the fluence rate
can affect E. coli inactivation (Bowker et al., 2011; Harm, 1980;
Sommer et al., 1998). As previously discussed, although UV-LEDs
are usually placed very close to the water sample for UV exposure
(Fig. 1), the fluence rate is still much lower than that of UV lamps

Table 2
UV sensitivity of microorganisms for UVC-LEDs and low-pressure UV lamps.

Microorganism Wavelength (nm) UV source k (cm2/mJ) Reference

4X174 255 UV-LEDs 0.578 Aoyagi et al., 2011

4X174 280 UV-LEDs 0.360 Aoyagi et al., 2011
4X174 254 UV lamps 0.396 Hijnen et al., 2006
E. coli 11229 255 UV-LEDs 0.300 Bowker et al., 2011
E. coli K12 29425 265 UV-LEDs 0.170 Chatterley and Linden, 2010
E. coli K12 IFO 3301 265 UV-LEDs 0.370 Oguma et al., 2013
E. coli 11229 275 UV-LEDs 0.422 Bowker et al., 2011
E. coli K12 IFO 3301 280 UV-LEDs 0.290 Oguma et al., 2013
E. coli 254 UV lamps 0.506 Hijnen et al., 2006
T7 255 UV-LEDs 0.195 Bowker et al., 2011
T7 275 UV-LEDs 0.235 Bowker et al., 2011
T7 254 UV lamps 0.232 Hijnen et al., 2006
Bacillus subtilis 250 UV-LEDs 0.051 Morris, 2012
Bacillus subtilis 269 UV-LEDs 0.148 Wurtele et al., 2011
Bacillus subtilis 282 UV-LEDs 0.120 Wurtele et al., 2011
Bacillus subtilis 254 UV lamps 0.059 Hijnen et al., 2006
Qb 255 UV-LEDs 0.080 Aoyagi et al., 2011
Qb 280 UV-LEDs 0.035 Aoyagi et al., 2011
Qb 254 UV lamps 0.084 Hijnen et al., 2006
MS2 255 UV-LEDs 0.078 Aoyagi et al., 2011
MS2 255 UV-LEDs 0.038 Bowker et al., 2011
MS2 275 UV-LEDs 0.035 Bowker et al., 2011
MS2 280 UV-LEDs 0.033 Aoyagi et al., 2011
MS2 254 UV lamps 0.055 Hijnen et al., 2006
346 K. Song et al. / Water Research 94 (2016) 341e349
due to the low output power of current UV-LEDs, which could be an
influencing factor. Bowker et al. (2011) compared a 254-nm UV
lamp with 275-nm and 255-nm UV-LEDs for E. coli inactivation. The
fluence rates were 0.34, 0.094e0.11, and 0.049e0.060 mW/cm2,
respectively. The results showed that log inactivation for the same
UV dose increased concurrently with the increase in fluence rate:
from the 255-nm UV-LED to the 275-nm UV-LED to the 254-nm UV
lamp. These results support the hypothesis that the combination of
a higher fluence rate and shorter exposure time may result in
higher E. coli inactivation as previously discussed, which could
explain the influence of different UV sources.
Bacteriophage T7 was found to have k values of 0.195 and
0.235 cm2/mJ by 255-nm and 275-nm UV-LEDs, respectively
(Bowker et al., 2011), which were slightly lower than those of E. coli,
indicating that T7 was also sensitive to UVC radiation. The k value
for 275 nm was slightly higher than that of 255 nm, suggesting that
T7 was more sensitive to 275 nm than to 255 nm, in agreement
with the action spectrum of T7 bacteriophage with a peak around
270 nm (Ronto et al., 1992; Beck et al., 2015). The k value of a UV
lamp for T7 was 0.232 cm2/mJ, which was close to that found with
the 275-nm UV-LED but far from the 255-nm UV-LED, implying
that T7 may also be sensitive to fluence rate, as previously
discussed.
Bacillus subtilis spores were inactivated in two studies (Morris,
2012; Wurtele et al., 2011) by UV-LEDs with three different wave-
lengths (250, 269, and 282 nm). The k values for 269 nm and
282 nm were 0.148 and 0.120 cm2/mJ, respectively, but the k value
decreased dramatically to 0.051 cm2/mJ for 250 nm. These results
seemed to be related to the action spectrum of B. subtilis spores,
which have a peak around 265 nm and a trough around 240 nm
(Mamane-Gravetz et al., 2005). The k value for B. subtilis spores by a
UV lamp was 0.059 cm2/mJ, which was similar to that of a 250-nm
UV-LED, showing a consistency in the UV sensitivity of B. subtilis
spores between the two different UV sources.
Bacteriophage Qb inactivation by 255-nm and 288-nm UV-LEDs
resulted in k values of 0.080 and 0.035 cm2/mJ, respectively (Aoyagi
et al., 2011), which were lower than most of the microorganisms
listed in Table 2, indicating that bacteriophage Qb was highly
resistant to UVC radiation. The k value at 255 nm was significantly
higher than that at 280 nm, which agreed well with the action
spectrum of Qb having a peak between 260 and 265 nm (Beck et al.,
2015). The k value for the 255-nm UV-LED agreed well with the
values from low-pressure UV lamp inactivation studies for Qb
(0.084 cm2/mJ), implying that Qb had the same UV sensitivity at
different UV sources.
Bacteriophage MS2 inactivation was reported in two studies
(Aoyagi et al., 2011; Bowker et al., 2011) using UV-LEDs at three
different wavelengths (255, 275, and 280 nm). The k values for
280 nm from the former study and 255, 275 nm from the latter
study were close (0.033, 0.038, and 0.035 cm2/mJ, respectively) and
among the lowest of all the microorganisms inactivated by UVC
radiation, indicating that MS2 was highly resistant to UVC radia-
tion. The k value for 255 nm from the former study was 0.078 cm2/
mJ, which was higher than the other values. This result matched the
action spectrum of MS2, which has a peak around 260 nm
(Mamane-Gravetz et al., 2005; Beck et al., 2015). Although these
two studies showed different k values for the 255-nm UV-LED on
MS2 inactivation, the average was 0.058 cm2/mJ, which was close to
that for MS2 inactivation by conventional UV lamps (0.055 cm2/mJ).
This finding suggested that MS2 might also be insensitive to
different fluence rates from UV-LEDs and conventional UV lamps.
These studies have demonstrated the diversity of UV responses
for different microorganisms to various wavelengths. Because UV
response largely depends on the indicator microorganisms and
applied wavelength by UV-LEDs, a particular wavelength could be
much more effective for a specific microorganism. Unlike UV lamps,
UV-LEDs can be designed to emit a particular wavelength that
targets a specific pathogen of concern. In addition, the selected
wavelengths could be combined by applying various UV-LEDs to
produce a synergistic effect (Taghipour, 2013). Therefore, a
comprehensive investigation of microorganism responses to UV-
LEDs with different wavelengths and wavelength combinations is
essential to take full advantage of UV-LEDs for water disinfection.
Additionally, from the comparison with conventional UV mer-
cury lamps at 254 nm, some microorganisms showed different UV
sensitivities to different UV sources at the same wavelength, indi-
cating that the characteristics and performance of UV sources, such
as the fluence rate, may play important roles in microorganism
inactivation. Microorganisms were previously shown to exhibit
higher UV sensitivities to medium-pressure mercury lamps than to
low-pressure mercury lamps (Hijnen et al., 2006). However, few
studies compare UV-LEDs and UV lamps. Thus, more studies are
required to better understand how different UV sources affect
microorganism inactivation using UV-LEDs compared with UV
lamps.
4. Time-response data on UV-LED disinfection
Some studies reported time-response data for microorganism
inactivation by UV-LEDs instead of doseeresponse data (Table 3).
The UV exposure times ranged from 30 s to 2 h, and log inactivation
ranged from 2 to 7; both results showed a wide range and seemed
incomparable. Because no quantitative data on UV dose were
provided in these studies, it was difficult to analyze and compare
the effectiveness of UV-LED inactivation. The difficulty was most
likely caused by the lack of an established standard method for the
measurement of UV dose associated with the newly emerging UV-
LEDs, which are substantially different from UV mercury lamps as
discussed previously. This issue demonstrates the significance of
UV-LED characterization and the need for standard protocols for
UV-LED microorganism inactivation and water disinfection.
5. Mechanism of inactivation
Many studies have discussed the mechanism of microorganism
inactivation by UV-LEDs, and various mechanisms have been pro-
posed. Much like the conventional mercury-based UV lamps, UVC
radiation has proven to be efficient for microorganism inactivation,
and most studies on UV-LED disinfection used UVC-LEDs. UVC ra-
diation is believed to have direct germicidal effects by acting
directly on the DNA of microorganisms, leading to the formation of
pyrimidine dimers and preventing them from reproducing without
intermediate steps (Chatterley and Linden, 2010; Chevremont et al.,
2012a, 2012b; Hamamoto et al., 2007). Because DNA mainly ab-
sorbs UV radiation from 200 to 300 nm with an absorbance peak
around 260 nm (Wurtele et al., 2011), UVC-LEDs, especially those
with the wavelengths around 260 nm, are the most efficient for
microorganism inactivation (Table 1). However, direct damage to
DNA might be reparable by DNA-repair mechanisms, such as
photo-reactivation and dark repair (Oguma et al., 2001, 2013; Sanz
et al., 2007). Since DNA repair is undesirable for microorganism
inactivation, it is necessary to weaken or prevent the repair. DNA
repair might be prevented by damaging the repair enzymes, which
are proposed to be more vulnerable to high UV intensities (Sommer
et al., 1998). Moreover, the absorption spectrum of protein has a
peak around 280 nm, which might help damage repair enzymes
and prevent DNA repair (Kalisvaart, 2004).
Although UVA radiation is poorly absorbed by DNA and is less
efficient in inducing damage on DNA (Sinha and Hader, 2002), it
still has the ability to inactivate microorganisms (Kalisvaart, 2004).
 K. Song et al. / Water Research 94 (2016) 341e349 347

Table 3
Summary of time-response data for water disinfection by UV-LEDs.

Wavelength (nm) Microorganism Disinfection medium Exposure time Log inactivation Reference

254 E. coli Water 30 s 3.5 Chevremont et al., 2012b

260 E. coli Water 50 min 2.5 Nelson et al., 2013
269 E. coli Water 5 min 3e4 Vilhunen et al., 2009
276 E. coli Water 5 min 3e4 Vilhunen et al., 2009
280 E. coli Water 30 s 7 Chevremont et al., 2012b
365 E. coli Water 30 s 2.7 Chevremont et al., 2012b
365(Pulse) E. coli Biofilm 1h 3 Li et al., 2010
365(Pulse) Candida albicans Biofilm 1h 3 Li et al., 2010
365 Vibrio parahaemolyticus Water 6 min 1 Nakahashi et al., 2014
375 E. coli Biofilm 2h 2 Hwang et al., 2013
375 Streptococcus mutans Water 15 min 4 Kim et al., 2007
405 E. coli Water 30 s 3.3 Chevremont et al., 2012b
254/365 E. coli Water 30 s 7 Chevremont et al., 2012b
280/365 E. coli Water 30 s 7 Chevremont et al., 2012b
280/405 E. coli Water 30 s 7 Chevremont et al., 2012b

The current commercially available UVA-LEDs have much higher
output power and energy efficiency than UVC-LEDs (Aoyagi et al.,
2011; Harris et al., 2013; Muramoto et al., 2014), resulting in
many studies on the application of UVA-LEDs for microorganism
inactivation. The inactivation mechanism of UVA radiation has not
been studied as widely as UVC radiation because the frequently
used UV mercury lamps can only emit UVC radiation at 254 nm. The
main mechanism of UVA inactivation involves an indirect effect by
reactive intermediates and oxidative damage to DNA and other
cellular components (Chatterley and Linden, 2010; Chevremont
et al., 2012a, 2012b; Hamamoto et al., 2007; Hwang et al., 2013).
The reactive intermediates are proposed to be reactive oxygen
species (ROS), which are created by UVA radiation via photooxi-
dation of oxygen. Studies showed that addition of mannitol and
catalase significantly protected microorganisms from 365-nm
UVA-LEDs radiation by scavenging hydroxyl radicals (OH) and
hydrogen peroxide (H2O2), respectively. Therefore, hydroxyl radi-
cals and hydrogen peroxide are believed to be the major reactive
oxygen species involved in UVA disinfection (Hamamoto et al.,
2007; Li et al., 2010). These reactive intermediates induce oxida-
tive damage to DNA, proteins, and cell membranes and cause
growth delay (Eisenstark, 1987; Oppezzo and Pizarro, 2001;
Pizarro, 1995; Pizarro and Orce, 1988; Ramabhadran and Jagger,
1976; Sinha and Hader, 2002). This process takes more time than
the direct damage produced by UVC radiation (Chatterley and
Linden, 2010). Although the indirect damage by UVA radiation is
less efficient than the direct damage by UVC radiation for micro-
organism inactivation, the damage by UVA is believed to be irrep-
arable, whereas the damage by UVC is reparable through DNA-
repair mechanisms (Oguma et al., 2013; Xiong and Hu, 2013). UV
damage by low-pressure mercury lamps, which emit UVC radiation
at 254 nm, can be repaired relatively easily, but UV damage induced
by medium-pressure mercury lamps, which produce UVC and UVA
radiation together, is difficult to repair (Oguma et al., 2002; Zimmer
and Slawson, 2002). Therefore, the prevention of microorganism
self-repair would be an advantage of UVA radiation for microor-
ganism inactivation. Furthermore, UVA radiation has higher
penetrability and can penetrate farther into the solution for a better
disinfection of turbid water and wastewater (Chevremont et al.,
2012a; Mori et al., 2007).
UVA radiation alone is less efficient for disinfection, but UVA
radiation coupled with TiO2 could efficiently produce reactive ox-
ygen species for microorganism inactivation (Marugan et al., 2010).
Because UVA-LEDs have high output power, they are desirable for
photocatalytic disinfection with TiO2. An interesting phenomenon
called “residual disinfecting effect” was reported, in which further
inactivation of E. coli was observed after a photocatalytic process
using a combination of UVA-LEDs and TiO2 (Xiong and Hu, 2013).
The mechanism for residual disinfecting effect was proposed to be
the cumulative damage of cellular components by reactive oxygen
species or stable oxidants, such as H2O2, which could prevent the
reproduction of damaged microorganisms (Pablos et al., 2013;
Rincon and Pulgarin, 2004, 2007; Shang et al., 2009).
Many studies have been conducted on microorganism inacti-
vation mechanisms using different wavelengths, including UVC and
UVA, with a focus on the effect on DNA damage. However, there is
little information in the literature on DNA damage and inactivation
mechanisms using a combination of different UV wavelengths
(Nakahashi et al., 2014). Because a few studies have reported the
synergistic effect of combining particular wavelengths
(Chevremont et al., 2012a; Nakahashi et al., 2014), identifying the
mechanisms is essential. Chevremont et al. (2012a) argued that
coupled wavelengths combined two UV properties: UVC induces
direct damage on DNA, but such DNA damage can be repaired by
enzyme photolyase, whereas the oxidative damage to bacterial
membranes by UVA cannot be repaired. The research on an
oxidative DNA product, 8-hydroxy-20 -deoxyguanosine (8-OHdG),
which was induced by UVA alone, and a thymine dimer, cyclo-
butane pyrimidine dimers (CPDs), which was induced by UVC
alone, suggested that the combination of UVA and UVC suppressed
one or more recovery systems for DNA damage, such as CPDs, and
oxidative stress from UVA may play a key role in this synergistic
effect (Nakahashi et al., 2014). It is proposed that the coupled
wavelengths of UVA and UVC may also reduce reactivation after
exposure due to the combined effects of two types of UV wave-
lengths (Chevremont et al., 2012a).
The mechanism of pulsed UV illumination by xenon lamps with
high energy and a broad spectrum has been widely studied for
inactivation and food decontamination. Yet, its effect is not well
understood, and there is little research on pulsed illumination by
UV-LEDs. Pulsed UV illumination has more instantaneous energy
than continuous UV illumination (Li et al., 2010), and additional
inactivation mechanisms have been proposed, including photo-
chemical, photothermal, and photophysical effects (Elmnasser
et al., 2007). Other than DNA damage by UV, it is believed that
pulsed UV treatment can prevent DNA repair due to inactivation of
the DNA-repair system and other enzymatic functions (Elmnasser
et al., 2007). The UV pulse with more instantaneous energy may
lead to localized overheating and membrane destruction
(Krishnamurthy et al., 2007). Furthermore, the repeated and con-
stant disturbance from high intensity pulses could induce cell
structure damage such as cell wall rupture, membrane damage, and
cellular content leakage (Krishnamurthy et al., 2010). As a result of
these additional effects, pulsed UV illumination is reported to be 4
348 K. Song et al. / Water Research 94 (2016) 341e349
to 6 times faster than continuous UV illumination for equivalent
inactivation levels (Fine and Gervais, 2004). These proposed
mechanisms are mainly based on studies of xenon lamps pulsed UV
illumination with high energy and a broad spectrum. Due to the
different pulses generated by xenon lamps and UV-LEDs, the
applicability of these mechanisms for UV-LED pulsed illumination
still needs further examination.
6. Future research
The research on water disinfection by newly emerging UV-LEDs
is limited. Because UV-LEDs are believed to be a promising alter-
native to traditional UV mercury lamps (Muramoto et al., 2014),
more research is required to better understand the application of
UV-LEDs for water disinfection. Further, some unique features of
UV-LEDs such as wavelength diversity and pulsed illumination and
their effects on microorganism inactivation must be explored and
understood. Given that UV-LEDs are compact and the radiation
patterns, such as emission wavelength, viewing angle, and radia-
tion distribution, are adjustable, they can enable creative reactor
designs through the optimization of flow and radiation distribu-
tion, as well as reactor geometry and kinetics (Taghipour, 2013). The
following are a few suggested areas for further investigation:
1) UV-LED characterization and a standard protocol for microor-
ganism inactivation are necessary to obtain reliable quantitative
information on the effectiveness of UV-LEDs. Specifically, a
standard method for UV-dose determination of UV-LEDs is
essential for accurate and reliable UV doseeresponse data,
which can be compared among different studies.
2) Multiple wavelengths and pulsed illumination by UV-LEDs can
have significant impacts on inactivation effectiveness, but more
studies are required for the fundamental understanding of these
phenomena, as well as determining the optimal condition.
3) The additional inactivation mechanisms by different wave-
length combinations and pulsed illumination require further
investigation, which could lead to the design of more efficient
disinfection systems using tailored combinations of selected
wavelengths and pulsation modes by UV-LEDs.
4) Research on reactor designs for UV-LED water disinfection sys-
tem is highly encouraged for the practical application of this
technology. The unique characteristics of UV-LEDs compared to
traditional UV mercury lamps, such as compactness, portability,
robustness, wavelength diversity, and pulsed illumination,
provide flexible and diverse options for novel reactor designs,
which could also open the door to new applications of UV-LED
reactors.
7. Conclusions
Newly emerging UV-LEDs provide a promising alternative for
water disinfection due to many advantages over traditional mer-
cury lamps. The comparison of microorganism response to UV-
LEDs and conventional UV lamps reveals that some microorgan-
isms may be sensitive to different UV sources, likely due to the
difference in the UV source radiation patterns and the fluence rates.
Inactivation studies of several microorganisms using UV-LED of the
same wavelengths and comparable fluence rates, however, still
show considerable discrepancies among published results. The
inconsistent and incomparable reported results on water disinfec-
tion by UV-LEDs along with the substantial differences between
UV-LEDs and UV lamps highlight the importance and necessity of
having a standard protocol for UV-LED microbial inactivation
studies, especially a standard method for UV-dose determination
using UV-LEDs.
The unique characteristics of UV-LEDs, particularly multiple
wavelengths and pulsed illumination, could improve inactivation
effectiveness under optimal conditions. However, the influencing
factors and mechanisms involved need to be further investigated
for a more efficient application of UV-LEDs. The special features of
UV-LEDs as small-point UV sources with adjustable radiation pat-
terns provide great flexibility for novel reactor designs and new
applications. The design of new UV-LED reactors, however, has to
be performed taking into account three phenomena: the reactor
hydrodynamics, radiation distribution, and kinetics. Each of these
phenomena may be implemented with a higher degree of freedom
using UV-LEDs as the radiation source compared to UV lamps.
Acknowledgements
The authors are grateful to the Natural Science and Engineering
Research Council (NSERC) of Canada for financial support. K. Song is
also grateful for a scholarship from the China Scholarship Council
(CSC).
References
Allied Analytics LLP, 2014. Global UV disinfection equipment market e size, share,
global trends, company profiles, analysis, segmentation and forecast,
2013e2020. Res. Mark. http://www.researchandmarkets.com/reports/3066124/
.
Aoyagi, Y., Takeuchi, M., Yoshida, K., Kurouchi, M., Yasui, N., Kamiko, N., Araki, T.,
Nanishi, Y., 2011. Inactivation of bacterial viruses in water using deep ultraviolet
semiconductor light-emitting diode. J. Environ. Eng. ASCE 137 (12), 1215e1218.
Autin, O., Romelot, C., Rust, L., Hart, J., Jarvis, P., MacAdam, J., Parsons, S.A.,
Jefferson, B., 2013. Evaluation of a UV-light emitting diodes unit for the removal
of micropollutants in water for low energy advanced oxidation processes.
Chemosphere 92 (6), 745e751.
Bak, J., Ladefoged, S.D., Tvede, M., Begovic, T., Gregersen, A., 2010. Disinfection of
Pseudomonas aeruginosa biofilm contaminated tube lumens with ultraviolet C
light emitting diodes. Biofouling 26 (1), 31e38.
Beck, S.E., Wright, H.B., Hargy, T.M., Larason, T.C., Linden, K.G., 2015. Action spectra
for validation of pathogen disinfection in medium-pressure ultraviolet (UV)
systems. Water Res. 70, 27e37.
Blatchley, E.R., 1997. Numerical modelling of UV intensity: application to
collimated-beam reactors and continuous-flow systems. Water Res. 31 (9),
2205e2218.
Bohrerova, Z., Shemer, H., Lantis, R., Impellitteri, C.A., Linden, K.G., 2008. Compar-
ative disinfection efficiency of pulsed and continuous-wave UV irradiation
technologies. Water Res. 42 (12), 2975e2982.
Bolton, J.R., Cotton, C.A., 2008. The Ultraviolet Disinfection Handbook. American
Water Works Association, Denver.
Bolton, J.R., Linden, K.G., 2003. Standardization of methods for fluence (UV dose)
determination in bench-scale UV experiments. J. Environ. Eng. ASCE 129 (3),
209e215.
Bowker, C., Sain, A., Shatalov, M., Ducoste, J., 2011. Microbial UV fluence-response
assessment using a novel UV-LED collimated beam system. Water Res. 45 (5),
2011e2019.
Brownell, S.A., Chakrabarti, A.R., Kaser, F.M., Connelly, L.G., Peletz, R.L., Reygadas, F.,
Lang, M.J., Kammen, D.M., Nelson, K.L., 2008. Assessment of a low-cost, point-
of-use, ultraviolet water disinfection technology. J. Water Health 6 (1), 53e65.
Chatterley, C., Linden, K., 2010. Demonstration and evaluation of germicidal UV-
LEDs for point-of-use water disinfection. J. Water Health 8 (3), 479e486.
Chen, R.Z., Craik, S.A., Bolton, J.R., 2009. Comparison of the action spectra and
relative DNA absorbance spectra of microorganisms: information important for
the determination of germicidal fluence (UV dose) in an ultraviolet disinfection
of water. Water Res. 43 (20), 5087e5096.
Chevremont, A.C., Boudenne, J.L., Coulomb, B., Farnet, A.M., 2013a. Fate of carba-
mazepine and anthracene in soils watered with UV-LED treated wastewaters.
Water Res. 47 (17), 6574e6584.
Chevremont, A.C., Boudenne, J.L., Coulomb, B., Farnet, A.M., 2013b. Impact of wa-
tering with UV-LED-treated wastewater on microbial and physico-chemical
parameters of soil. Water Res. 47 (6), 1971e1982.
Chevremont, A.C., Farnet, A.M., Coulomb, B., Boudenne, J.L., 2012a. Effect of coupled
UV-A and UV-C LEDs on both microbiological and chemical pollution of urban
wastewaters. Sci. Total Environ. 426, 304e310.
Chevremont, A.C., Farnet, A.M., Sergent, M., Coulomb, B., Boudenne, J.L., 2012b.
Multivariate optimization of fecal bioindicator inactivation by coupling UV-A
and UV-C LEDs. Desalination 285, 219e225.
Close, J., Ip, J., Lam, K.H., 2006. Water recycling with PV-powered UV-LED disin-
fection. Renew. Energy 31 (11), 1657e1664.
Eisenstark, A., 1987. Mutagenic and lethal effects of near-ultraviolet radiation (290-
 K. Song et al. / Water Research 94 (2016) 341e349
400 nm) on bacteria and phage. Environ. Mol. Mutagen 10 (3), 317e337.
Elmnasser, N., Guillou, S., Leroi, F., Orange, N., Bakhrouf, A., Federighi, M., 2007.
Pulsed-light system as a novel food decontamination technology: a review. Can.
J. Microbiol. 53 (7), 813e821.
Fine, F., Gervais, P., 2004. Efficiency of pulsed UV light for microbial decontamina-
tion of food powders. J. Food Prot. 67 (4), 787e792.
Gomez-Lopez, V.M., Ragaert, P., Debevere, J., Devlieghere, F., 2007. Pulsed light for
food decontamination: a review. Trends Food Sci. Technol. 18 (9), 464e473.
Hamamoto, A., Mori, M., Takahashi, A., Nakano, M., Wakikawa, N., Akutagawa, M.,
Ikehara, T., Nakaya, Y., Kinouchi, Y., 2007. New water disinfection system using
UVA light-emitting diodes. J. Appl. Microbiol. 103 (6), 2291e2298.
Harm, W., 1980. Biological Effects of Ultraviolet Radiation. Cambridge University
Press, New York, pp. 127e130.
Harris, T.R., Pagan, J., Batoni, P., 2013. Optical and fluidic co-design of a UV-LED
water disinfection chamber. ECS Trans. 45 (17), 11e18.
Hatami, H., 2013. Importance of water and water-borne diseases: on the occasion of
the world water day (March 22, 2013). Int. J. Prev. Med. 4 (3), 243.
Hijnen, W.A.M., Beerendonk, E.F., Medema, G.J., 2006. Inactivation credit of UV ra-
diation for viruses, bacteria and protozoan (oo)cysts in water: a review. Water
Res. 40 (1), 3e22.
Hwang, K.S., Jeon, Y.S., Choi, T.I., Hwangbo, S., 2013. Combination of light emitting
diode at 375 nm and photo-reactive TiO2 layer prepared by electrostatic
spraying for sterilization. J. Electr. Eng. Technol. 8 (5), 1169e1174.
Ibrahim, M.A.S., MacAdam, J., Autin, O., Jefferson, B., 2014. Evaluating the impact of
LED bulb development on the economic viability of ultraviolet technology for
disinfection. Environ. Technol. 35 (4), 400e406.
IUVA, 2015. Proposed testing protocol for measurement of UV-C LED lamp output.
IUVA NEWS 17 (2), 7.
Kalisvaart, B.F., 2004. Re-use of wastewater: preventing the recovery of pathogens
by using medium-pressure UV lamp technology. Water Sci. Technol. 50 (6),
337e344.
Khan, M.A., Shatalov, M., Maruska, H.P., Wang, H.M., Kuokstis, E., 2005. III-nitride UV
devices. Jpn. J. Appl. Phys. 44 (10), 7191e7206.
Kim, B.H., Kim, D., Cho, D.L., Lim, S.H., Yoo, S.Y., Kook, J.K., Cho, Y.I., Ohk, S.H.,
Ko, Y.M., 2007. Sterilization effects of a TiO2 photocatalytic film against a
Streptococcus mutans culture. Biotechnol. Bioprocess Eng. 12 (2), 136e139.
Krishnamurthy, K., Demirci, A., Irudayaraj, J.M., 2007. Inactivation of Staphylococcus
aureus in milk using flow-through pulsed UV-Light treatment system. J. Food
Sci. 72 (7), M233eM239.
Krishnamurthy, K., Tewari, J.C., Irudayaraj, J., Demirci, A., 2010. Microscopic and
spectroscopic evaluation of inactivation of Staphylococcus aureus by pulsed UV
light and infrared heating. Food Bioprocess Technol. 3 (1), 93e104.
Kuo, J., Chen, C.L., Nellor, M., 2003. Standardized collimated beam testing protocol
for water/wastewater ultraviolet disinfection. J. Environ. Eng. ASCE 129 (8),
774e779.
Li, J., Hirota, K., Yumoto, H., Matsuo, T., Miyake, Y., Ichikawa, T., 2010. Enhanced
germicidal effects of pulsed UV-LED irradiation on biofilms. J. Appl. Microbiol.
109 (6), 2183e2190.
Linden, K.G., Shin, G., Sobsey, M.D., 2001. Comparative effectiveness of UV wave-
lengths for the inactivation of Cryptosporidium parvum oocysts in water. Water
Sci. Technol. 43 (12), 171e174.
Malley, J.P., Ballester, N.A., Linden, K.G., Mofidi, A., Bolton, J.R., Margolin, A.B.,
Crozes, G., Cushing, B., Mackey, E., Lane, J.M., Janex, M.L., 2004. Inactivation of
Pathogens with Innovative UV Technologies. American Research Foundation
and American Water Works Association.
Mamane-Gravetz, H., Linden, K.G., Cabaj, A., Sommer, R., 2005. Spectral sensitivity
of Bacillus subtilis spores and MS2 coliphage for validation testing of ultraviolet
reactors for water disinfection. Environ. Sci. Technol. 39 (20), 7845e7852.
Marugan, J., van Grieken, R., Pablos, C., Sordo, C., 2010. Analogies and differences
between photocatalytic oxidation of chemicals and photocatalytic inactivation
of microorganisms. Water Res. 44 (3), 789e796.
Mori, M., Hamamoto, A., Takahashi, A., Nakano, M., Wakikawa, N., Tachibana, S.,
Ikehara, T., Nakaya, Y., Akutagawa, M., Kinouchi, Y., 2007. Development of a new
water sterilization device with a 365 nm UV-LED. Med. Biol. Eng. Comput. 45
(12), 1237e1241.
Morris, J.P., 2012. Disinfection of Bacillus subtilis Spores Using Ultraviolet Light
Emitting Diodes (MS thesis).
Muller, J., 2011. Seeing the Light: the Benefits of UV Water Treatment. Water Online.
http://www.wateronline.com/doc/seeing-the-light-the-benefits-of-uv-0001.
Muramoto, Y., Kimura, M., Nouda, S., 2014. Development and future of ultraviolet
light-emitting diodes: UV-LED will replace the UV lamp. Semicond. Sci. Technol.
29 (8).
Murata, Y., Osakabe, M., 2013. The Bunsen-Roscoe reciprocity law in ultraviolet-B-
induced mortality of the two-spotted spider mite Tetranychus urticae. J. Insect
Physiol. 59 (3), 241e247.
Nakahashi, M., Mawatari, K., Hirata, A., Maetani, M., Shimohata, T., Uebanso, T.,
Hamada, Y., Akutagawa, M., Kinouchi, Y., Takahashi, A., 2014. Simultaneous
irradiation with different wavelengths of ultraviolet light has synergistic
bactericidal effect on Vibrio parahaemolyticus. Photochem. Photobiol. 90 (6),
349
1397e1403.
Nelson, K.Y., McMartin, D.W., Yost, C.K., Runtz, K.J., Ono, T., 2013. Point-of-use water
disinfection using UV light-emitting diodes to reduce bacterial contamination.
Environ. Sci. Pollut. Res. 20 (8), 5441e5448.
Oguma, K., Katayama, H., Mitani, H., Morita, S., Hirata, T., Ohgaki, S., 2001. Deter-
mination of pyrimidine dimers in Escherichia coli and Cryptosporidium parvum
during UV light inactivation, photoreactivation, and dark repair. Appl. Environ.
Microbiol. 67 (10), 4630e4637.
Oguma, K., Katayama, H., Ohgaki, S., 2002. Photoreactivation of Escherichia coli after
low- or medium-pressure UV disinfection determined by an endonuclease
sensitive site assay. Appl. Environ. Microbiol. 68 (12), 6029e6035.
Oguma, K., Kita, R., Sakai, H., Murakami, M., Takizawa, S., 2013. Application of UV
light emitting diodes to batch and flow-through water disinfection systems.
Desalination 328, 24e30.
Oms-Oliu, G., Martin-Belloso, O., Soliva-Fortuny, R., 2010. Pulsed light treatments
for food preservation: a review. Food Bioprocess Technol. 3 (1), 13e23.
Oppezzo, O.J., Pizarro, R.A., 2001. Sublethal effects of ultraviolet A radiation on
Enterobacter cloacae. J. Photochem. Photobiol. B Biol. 62 (3), 158e165.
Pablos, C., Marugan, J., van Grieken, R., Serrano, E., 2013. Emerging micropollutant
oxidation during disinfection processes using UV-C, UV-C/H2O2, UV-A/TiO2 and
UV-A/TiO2/H2O2. Water Res. 47 (3), 1237e1245.
Pizarro, R.A., 1995. UV-A oxidative damage modified by environmental conditions
in Escherichia coli. Int. J. Radiat. Biol. 68 (3), 293e299.
Pizarro, R.A., Orce, L.V., 1988. Membrane damage and recovery associated with
growth delay induced by near-UV radiation in Escherichia coli K-12. Photochem.
Photobiol. 47 (3), 391e397.
Ramabhadran, T.V., Jagger, J., 1976. Mechanism of growth delay induced in Escher-
ichia coli by near ultraviolet radiation. Proc. Natl. Acad. Sci. U. S. A. 73 (1),
59e63.
Rincon, A.G., Pulgarin, C., 2004. Bactericidal action of illuminated TiO2 on pure
Escherichia coli and natural bacterial consortia: post-irradiation events in the
dark and assessment of the effective disinfection time. Appl. Catal. B Environ. 49
(2), 99e112.
Rincon, A.G., Pulgarin, C., 2007. Absence of E. coli regrowth after Fe3þ and TiO2 solar
photoassisted disinfection of water in CPC solar photoreactor. Catal. Today 124
(3e4), 204e214.
Ronto, G., Gaspar, S., Berces, A., 1992. Phage T7 in biological UV dose measurement.
J. Photochem. Photobiol. B Biol. 12 (3), 285e294.
Sanz, E.N., Davila, I.S., Balao, J.A.A., Alonso, J.M.Q., 2007. Modelling of reactivation
after UV disinfection: effect of UV-C dose on subsequent photoreactivation and
dark repair. Water Res. 41 (14), 3141e3151.
Shang, C., Cheung, L.M., Ho, C.M., Zeng, M.Z., 2009. Repression of photoreactivation
and dark repair of coliform bacteria by TiO2-modified UV-C disinfection. Appl.
Catal. B Environ. 89 (3e4), 536e542.
Sinha, R.P., Hader, D.P., 2002. UV-induced DNA damage and repair: a review. Pho-
tochem. Photobiol. Sci. 1 (4), 225e236.
Sommer, R., Haider, T., Cabaj, A., Pribil, W., Lhotsky, M., 1998. Time dose reciprocity
in UV disinfection of water. Water Sci. Technol. 38 (12), 145e150.
Sommer, R., Lhotsky, M., Haider, T., Cabaj, A., 2000. UV inactivation, liquid-holding
recovery, and photoreactivation of Escherichia coli O157 and other pathogenic
Escherichia coli strains in water. J. Food Prot. 63 (8), 1015e1020.
Taghipour, F., 2013. Is UV-LED the future of ultraviolet water purification? IUVA
NEWS 15 (4), 23e26.
Taniyasu, Y., Kasu, M., 2010. Improved emission efficiency of 210-nm deep-
ultraviolet aluminum nitride light-emitting diode. NTT Tech. Rev. 8 (8), 1e5.
Taniyasu, Y., Kasu, M., Makimoto, T., 2006a. An aluminium nitride light-emitting
diode with a wavelength of 210 nanometres. Nature 441 (7091), 325e328.
Taniyasu, Y., Kasu, M., Makimoto, T., 2006b. Aluminum nitride deep-ultraviolet
light-emitting diodes. NTT Tech. Rev. 4 (12), 54e58.
USEPA, 2006. Ultraviolet Disinfection Guidance Manual for the Final Long Term 2
Enhanced Surface Water Treatment Rule. Office of Water, Washington DC. EPA
815-R-06e007.
Vilhunen, S., Sarkka, H., Sillanpaa, M., 2009. Ultraviolet light-emitting diodes in
water disinfection. Environ. Sci. Pollut. Res. 16 (4), 439e442.
Wengraitis, S., McCubbin, P., Wade, M.M., Biggs, T.D., Hall, S., Williams, L.I.,
Zulich, A.W., 2013. Pulsed UV-C disinfection of Escherichia coli with light-
emitting diodes, emitted at various repetition rates and duty cycles. Photo-
chem. Photobiol. 89 (1), 127e131.
WHO, 2014. Progress on Drinking Water and Sanitation. http://apps.who.int/iris/
bitstream/10665/112727/1/9789241507240_eng.pdf.
Wurtele, M.A., Kolbe, T., Lipsz, M., Kulberg, A., Weyers, M., Kneissl, M., Jekel, M.,
2011. Application of GaN-based ultraviolet-C light emitting diodes e UV LEDs e
for water disinfection. Water Res. 45 (3), 1481e1489.
Xiong, P., Hu, J.Y., 2013. Inactivation/reactivation of antibiotic-resistant bacteria by a
novel UVA/LED/TiO2 system. Water Res. 47 (13), 4547e4555.
Zimmer, J.L., Slawson, R.M., 2002. Potential repair of Escherichia coli DNA following
exposure to UV radiation from both medium- and low-pressure UV sources
used in drinking water treatment. Appl. Environ. Microbiol. 68 (7), 3293e3299.