Content uploaded by Karl G Linden
Author content
All content in this area was uploaded by Karl G Linden
Content may be subject to copyright.
Demonstration and evaluation of germicidal UV-LEDs for
point-of-use water disinfection
Christie Chatterley and Karl Linden
ABSTRACT
Christie Chatterley
Karl Linden (corresponding author)
Department of Civil, Environmental, and
Architectural Engineering,
University of Colorado at Boulder,
Boulder, CO 80304,
USA
Tel.: 303-492-4798
Fax: 303-492-7317
E-mail: karl.linden@colorado.edu
Ultraviolet (UV) irradiation is a common disinfection option for water treatment in the developed
world. There are a few systems installed in developing countries for point-of-use treatment, but
the low-pressure mercury lamps currently used as the UV irradiation source have a number of
sustainability issues including a fragile envelope, a short lifetime of approximately one year,
and they contain toxic mercury. UV light emitting diodes (LEDs) may present solutions to many of
the sustainability issues presented by current UV systems. LEDs are small, efficient, have long
lifetimes, and do not contain mercury. Germicidal UV LEDs emitting at 265 nm were evaluated
for inactivation of E. coli in water and compared to conventional low-pressure UV lamps. Both
systems provided an equivalent level of treatment. A UV-LED prototype was developed and
evaluated as a proof-of-concept of this technology for a point-of-use disinfection option, and
the economics of UV-LEDs were evaluated.
Key words
|
household water treatment, light emitting diodes, ultraviolet light
INTRODUCTION
Diarrheal illnesses are one of the leading causes of
morbidity and mortality in developing countries (Pruss
2002;WHO 2002). In many analyses of interventions to
reduce diarrhea, “improved water quality” is shown to have
a lower effect than other interventions such as sanitation
and hygiene. However, these reviews focus upon source
water quality improvements rather than improvements at
point-of-use (Gundry et al. 2004). Fewtrell & Colford
(2004), showed the increased impact of treating water at
the household level compared to treating at the source. This
information has created an interest in household water
treatment technologies. A number of point-of-use technol-
ogies have been evaluated including boiling, biosand
filtration, chlorination, chlorination plus flocculation,
solar disinfection (SODIS), and ceramic filters (Sobsey
et al. 2008). Disinfection using ultraviolet (UV) radiation in
the UV-C range may be a more favorable option for many
applications. It does not utilize chemicals and disinfects at
much higher rates than SODIS which utilizes temperature
and radiation in the UV-A range.
UV disinfection is a well-established disinfection tech-
nology that has been used in centralized water and
wastewater facilities in developed countries for decades.
UV radiation inactivates bacteria, viruses, and protozoa,
with the benefits of no taste and odor issues, no known
disinfection byproducts (DBPs), no danger of overdosing,
relatively fast treatment rates compared to sand and
ceramic filters, and low-maintenance requirements. Over
the last ten years, small UV systems have become available,
including commercially available household systems and
the low-cost, locally manufactured UV-Tube system that
have become an appropriate treatment option for develop-
ing communities in a number of countries including
Mexico, Sri Lanka, and India (Brownell et al. 2008).
For developing communities, UV disinfection can be
an improvement over other treatment options, such as
doi: 10.2166/wh.2010.124
1QIWA Publishing 2010 Journal of Water and Health
|
inpress
|
2010
Uncorrected Proof
chemical disinfection, for many applications, but there are
sustainability issues that arise from current low-pressure
lamp systems in use. They use toxic mercury as the UV
radiation source and typically only last for 8,000 –10,000
hours, at which time communities are faced with a number
of issues: finding and paying for replacement lamps,
transporting these fragile glass and filament tubes, and
disposing of mercury contained in the used lamp in areas
that do not typically have a toxic waste disposal system
(US EPA 2006).
UV light emitting diodes (LEDs) may provide solutions
to many of the sustainability issues of UV mercury lamps.
They are small (5–9 mm diameter), and do not contain
glass, filament or mercury, aiding their transport and
disposal (Bettles et al. 2007). Warm-up time is not required
for LEDs, saving energy and allowing for intermittent use
and quick recovery from a power failure—important
characteristics for rural applications especially. LEDs are
replacing a number of light sources currently utilized today
including traffic lights and household lights. LEDs have an
excellent track record for lowering system costs through
energy savings, lower maintenance, and longer replacement
intervals. The average electrical-to-germicidal efficiency of
low-pressure UV mercury tube lamps is 35–38% (US EPA
2006). Visible LEDs can operate at 75% efficiency for ten
years (100,000 hours) (Bettles et al. 2007). Currently, the
efficiencies of UV-LEDs are less than 1% with lifetimes of
around 1,000 hours (Bettles et al. 2007;Gaska 2007).
Although research of this technology is still in its infancy,
improvements to UV-LEDs are expected to occur rapidly
following visible LED source trajectories, resulting in a high
efficiency, low power input.
The availability of specific output wavelengths using
UV-LEDs may also increase their inactivation efficacy.
UV-LEDs currently operate in the wavelength range of
247 –365 nm (Gaska 2007). Effective UV sources should
emit high intensities in the peak absorbance wavelengths of
DNA—the germicidal target of UV photons. However,
germicidal effectiveness as a function of wavelength can
vary for different microorganisms and may differ from the
DNA absorbance spectrum. Supplementing peak DNA
wavelengths with other UV emissions may provide a
synergistic disinfection effect, increasing the effectiveness
of UV inactivation of pathogens (Mamane-Gravetz et al.
2005;US EPA 2006;Linden et al. 2007). Low-pressure
lamps are monochromatic (254 nm) and some pathogens,
such as adenovirus, are not most effectively inactivated at
this wavelength. Medium Pressure lamps are polychromatic,
but peak intensities occur at set wavelengths based on the
emission properties of mercury. A distinct advantage over
conventional UV sources is that UV-LED systems can
incorporate an LED array of differing UV wavelengths,
maximizing their combined germicidal effect. This would
allow units to be custom designed based on the specific
pathogens of concern in source waters, or for a broad range
of pathogens under a single system.
Limited research has been conducted on the effective-
ness of UV-LEDs for water disinfection. Most of the data
available are for LEDs that emit light in the UVA range
(320 –400 nm), which is less efficient at disinfection than
light in the germicidal range of UVC (200–280 nm) since it
is poorly absorbed by DNA (Sinha & Ha¨ der 2002; ISO
21348 2007). UVA radiation inactivates microorganisms by
damaging proteins and producing hydroxyl and oxygen
radicals that can destroy cell membranes and other cellular
components (Sinha & Ha¨ der 2002). This process takes more
time than the damage produced by UV-C, which directly
effects the DNA of microorganisms by producing cyclobu-
tane thymine dimers, among other products, inactivating
them without intermediate steps (Grossweiner & Smith
1989). Hamamoto et al. (2007) demonstrated the ability of
UVA-LEDs at 365 nm to inactivate bacteria in water. They
found that E. coli DH5awere reduced by .5 log at a dose
of 315 J/cm
2
approximately 30,000 times higher dose than
required for UV 254 nm. Sandia National Laboratories
documented inactivation of E. coli with LEDs in the UVC
range at 270 nm (Crawford et al. 2005), and found
comparable inactivation to LP UV. Sensor Electronics
Technologies (SET) has also demonstrated inactivation of
E. coli B using 265 –310 nm UV-LEDs (Gaska 2007),
reporting wavelength dependent inactivation with inacti-
vation decreasing by more than 6 orders of magnitude from
265 nm to 310 nm.
Research objectives
The goal of this research was to evaluate the efficacy of
Ultraviolet Light Emitting Diode (UV-LED) technology for
2C. Chatterley and K. Linden
|
UV-LEDs for water disinfection Journal of Water and Health
|
inpress
|
2010
Uncorrected Proof
the development of point-of-use (POU) water disinfection
systems to improve public health in rural communities in a
sustainable, environmentally responsible manner. There are
a number of POU technologies available, but the appli-
cation of UV-LEDs as a disinfection source will provide an
additional technology to the POU toolbox that will enable
longer-life disinfection systems with low user input and very
low energy cost compared to current low-pressure mercury
lamps. This will improve public health by increasing system
reliability and decreasing maintenance needs.
Specifically, this research evaluated the use of UV-
LEDs at 265 nm for inactivation of E. coli in water through
the following objectives: (1) Compare the inactivation
efficiency of UV-LEDs at 265nm to conventional low-
pressure lamps (254 nm) for inactivation of E. coli, (2)
Evaluate a point-of-use UV-LED flow-through prototype,
and (3) Determine if UV-LEDs are a feasible option for
water treatment based on economics and current state of
the technology.
METHODS
Microbial methods
E. coli K12 (ATCC #29425) was used as an indicator
organism to compare the efficiency of the LP and UV-LED
systems, and to evaluate the UV-LED prototype. One
colony (to assure genetic homogeneity) was obtained from
a tryptic soy agar (TSA, Difco #236950) plate after 24 hours
of incubation at 378C and added to 10 mL of sterile tryptic
soy broth (TSB, Cellgro #61-412-RO) in a sterile 15 mL vial.
The vial was rapidly vortexed to break up the colony and
then the 10 mL solution was added to 90 mL of TSB in a
sterile 250 mL glass bottle with a sterile magnetic stir bar.
The stock solution was incubated at 378C on a stir-plate to
assure constant mixing and oxygen levels throughout the
stock. This solution was kept at 48C for less than 2 weeks to
inoculate future stock solutions. Purity was verified by
streak plating and visual observation on TSA.
A growth curve was developed based on the optical
density at 600 nm (OD600), measured every 30 minutes,
and cultured colonies to identify the log growth phase. Tests
were conducted at log growth phase. The E. coli were
washed three times in phosphate buffer solution (PBS) by
centrifuging and added to 194 mL of PBS to achieve a
concentration of approximately 10
6
CFU E. coli per mL for
batch irradiation testing. The ultraviolet absorbance was
adjusted for flow-through tests by varying the E. coli
preparation steps (washing).
After irradiation, each sample was successively vortexed
and serially diluted. E. coli concentrations were measured
using the spot plating method (Gaudy et al. 1962), which is
advantageous because four dilutions of five replicates each
can be read on one 100 mm diameter plate, which would
otherwise require twenty 60 mm diameter plates using the
vacuum filtration method. Once the spots had completely
dried, the plates were placed upside down in a 378C
incubator and incubated for 24 hours before colonies
were counted. Spots with 3 to 30 colonies were recorded
(CFU/0.01 mL). If two dilutions had results that fell into
this range, the lower dilution (more colonies in each spot)
was chosen.
Experimental set-up
Low-pressure (LP) UV lamps were housed in a UV
collimated beam apparatus (Bolton & Linden 2003).
A UV-LED batch irradiation system was designed with an
array of three UV-LEDs using a circuit wire-wrapped with
30 gauge wire, to an electronic Perfboard (a fiberglass board
with holes every 2.5 mm). A 150-ohm resistor was wired in
series with each LED to create 6 volts across each LED at
20 amps with a 9 volt input voltage from a power supply.
These values were within manufacturer specifications for
voltage and current. Socket pins were wire-wrapped to
the Perfboard to hold the LEDs in place for easy removal
and replacement. A flow-through prototype consisting of
a compact row of ten UV-LEDs was created using similar
electronics to the batch system. The LEDs were placed over a
6.5 mm £6.5 mm aluminum channel 1 mm above the water
surface, with a water depth of 7 mm. The 265 nm hemi-
spherical UV LEDs were purchased from Sensor Electronic
Technology, Inc (SET) (Columbia, South Carolina).
Irradiance measurements
Irradiance for the LP and LED light sources was measured
with a radiometer (International Light IL1400A, SEL
3C. Chatterley and K. Linden
|
UV-LEDs for water disinfection Journal of Water and Health
|
inpress
|
2010
Uncorrected Proof
240/TD detector) calibrated at 254 nm. The manufacturer’s
detector response curve was used to adjust the radiometer
reading to the 265 nm peak output wavelength of the LEDs.
The radiometer measurements were checked and corrobo-
rated using an iodide/iodate actinometer (Rahn et al. 2003).
The absolute irradiance and spectral output of each LED
was also evaluated using an Ocean Optics spectrometer
(USB 2000 þ, Dunadin, FL).
UV irradiation
All tests were completed within two hours and irradiated
samples were covered to minimize photoreactivation as
much as possible.
To benchmark the efficiency of LP UV, the collimated
beam apparatus was used to expose 40 mL portions of E.
coli spiked PBS at UV fluences ranging from 0 to 20 mJ/cm
2
in a sterile 50 mL glass crystallization dish (2.2cm diameter)
stirred with a sterile magnetic stir bar. The UV-LED devices
were evaluated by exposing 7 mL of E. coli spiked PBS in a
10 mL beaker (2.2 cm diameter) stirred with a sterile
magnetic stir bar to UV doses between 0 and 20 mJ/cm
2
.
The UV-LED prototype was evaluated by flowing E. coli
spiked PBS and E. coli spiked natural water (collected from
a local pond) through the system. Initial E. coli concen-
tration was tested by running the sample through the
prototype with the LEDs turned off. Log reduction of E. coli
was evaluated for multiple flow rates and multiple UV
absorbance values. The system was disinfected between
tests using a low concentration chlorine solution.
UV dose calculations
The average irradiance in the UV-exposed sample was
calculated according to Bolton and Linden (Bolton &
Linden 2003). A petri factor of 0.98 and 1 were determined
for the LP and LED systems, respectively, and a reflection
factor for water of 0.975 was used. The water factor
accounted for the UV absorbance of the water through
the sample water depth at 254 nm and 266 nm for the LP
and LED systems, respectively (measured with a spectro-
photometer, HACH DR 5000).
Irradiation time was controlled by a manual shutter for
LP tests and by turning on/off the LEDs. The LP lamps and
LEDs were allowed to warm-up for 10 minutes before tests
and the LEDs were turned off for a maximum of 10 seconds
while tests were being set-up, which did not significantly
affect the irradiance upon turning back on.
The E. coli colonies were averaged for the five spots and
converted to CFU/mL by multiplying by the dilution factor.
The log reduction (log N
0
/N) was calculated for each dose
based on the initial non-irradiated E. coli concentration, N
0
,
(CFU/mL) and the concentration of E. coli post-irradiation,
N (CFU/mL).
RESULTS AND DISCUSSION
Irradiance during warm-up time
The irradiance over time (0, 1, 2, 5, 10, and 20 minutes from
start) was measured to determine the warm-up time of both
the low-pressure (LP) lamps and the UV-LEDs. Over the
first 10 minutes after start-up, the irradiance of the UV-
LEDs decreases by about 7% and the irradiance of the LP
lamps increases by about 20%, after which time both
sources level out (Figure 1).
Seven LEDs from SET were tested including four flat
top 265 nm LEDs, one flat top 250 nm LED, one flat top
280 nm LED, and one hemispherical lens 280 nm LED. The
spectral irradiance of each LED was measured with a
spectrometer (Ocean Optics USB 2000 þ) as presented in
Figure 2. The UV-LEDs from SET had a broader bandwidth
than the monochromatic 253.7 nm produced by low-
pressure lamps. The full width at half maximum (FWHM),
measured across the spectral output at 50% of the peak
Figure 1
|
Warm-up time for UV-LEDs (B) versus Low Pressure Lamps (O).
4C. Chatterley and K. Linden
|
UV-LEDs for water disinfection Journal of Water and Health
|
inpress
|
2010
Uncorrected Proof
irradiance, was 11 nm for the 265 nm LED, slightly lower
than the manufacturer FWHM specification of 12 nm. The
broader emission spectra could have implications for
system designs, paricularly if very specific wavelengths
are desired.
Irradiance from a single 265 nm hemispherical lens
UV-LEDs was measured at distances from the source up to
four cm, to estimate the effect of distance on irradiance
changes, specifically for future modeling and estimating UV
dose for the prototype unit. The irradiance output from the
265 nm LEDs tested varied from 60 to 30 mW/cm
2
,at
distances of 0.5 to 4 cm as illustrated in Figure 3.
Inactivation of E. coli
Low-pressure versus UV-LEDs
Log reduction of E. coli K12 appears to be slightly improved
for the LED source at low doses and approximately the
same at higher doses, based on twenty three and eighteen
data points in duplicate for the LED and LP sources
respectively, as illustrated in Figure 4.
The relationship of log inactivation versus dose received
was modeled using a logarithmic regression. Based on
results of paired t-tests conducted over the log inactivation
data at each dose (2, 5, 10, 15, and 20 mJ/cm
2
) for low-
pressure versus LED sources, it can not be concluded that
the low-pressure and LED sources are statistically different
for the inactivation of E. coli K12 at a 95% confidence,
although there is a statistically significant difference at a
90% confidence level.
UV-LED flow-through prototype
The ten-LED prototype was evaluated using biodosimetry
with E. coli K12. The linear trendlines for log reduction with
varying UV absorbance values all have a similar slope
(within one log reduction per one liter per hour) and the
waters with lower UV absorbance values are disinfected to
the same level as waters with higher UV absorbance values
when lower flow rates were used (Figure 5). The inacti-
vation in natural water (UV absorbance of 0.259) was
similar to the PBS samples.
In order to compare the prototype to commercial
systems, the dose provided for a given flow rate in
mL/min and influent water UV transmittance (UVT) are
needed. The log reduction of E. coli for a given UV-LED
dose was calculated based on the logarithmic regression
model (log reduction ¼1.25 £Ln(Dose) 20.3665) on the
Figure 2
|
Irradiance and spectral emission from 250, 265, and 280 nm UV-LEDs.
Figure 3
|
Irradiance of one 265 nm LED as a function of distance from the LED.
Figure 4
|
Log reduction of E. coli K12 by irradiation from low-pressure lamps (254 nm)
and LEDs (265 nm). Data lines represent logarithmic regression and error
bars represent one standard deviation of the mean of quintuplicates.
5C. Chatterley and K. Linden
|
UV-LEDs for water disinfection Journal of Water and Health
|
inpress
|
2010
Uncorrected Proof
E. coli inactivation data presented in Figure 4. For a UVT of
88%, a dose of 10 mJ/cm
2
can be achieved at a flow rate of
14 mL/min. For a UVT of 74%, the flow rate required for a
dose of 10 mJ/cm
2
is 11.1 mL/min. Therefore, forty LEDs
(four rows of 10 LEDs in the same geometry as the
prototype) at a flow rate of 11.1 mL/min could provide a
dose of roughly 40 mJ/cm
2
, a typical dose for commercial
UV disinfection systems. Forty LEDs emitting at 265 nm
have a total rated output of 14.4 mW of power, therefore
1.3 mW are needed per mL/min flow rate to achieve a dose
of 40 mJ/cm
2
. This value is used to compare LED and LP
systems in Table 1. However, direct comparison is highly
dependent on reactor hydraulics and the comparison
presented is used only as a case study based on the LED
prototype evaluated. The power requirements for UV-LED
systems have the potential to be reduced by optimizing
reactor geometry and water depth. The use of UV-LEDs
may facilitate optimization through creative system design
options based on their small individual size.
Evaluation of current and future UV-LED technology
LEDs that emit light in the germicidal wavelength range are
a relatively new technology and current values for cost,
output power, and lifetime do not at present allow them to
be a viable option for the replacement of low-pressure
lamps used for drinking water disinfection, especially in
developing communities. Based on a household system that
needs to provide twenty liters per person per day for a
family of four (eighty liters per day total), and a dose of
40 mJ/cm
2
and 75% UVT, a comparison was conducted of
current UV-LEDs with current LP systems such as the
UV-Tube and the Sterilight (R-Can Environmental Inc.,
Guelph, Ontario, Canada) systems. The base case includes
current UV-LED specifications and assumes a constantly
running system. The comparison presented in Table 1
shows the much greater cost of UV-LEDs, both upfront
and over time since the lifetime is much lower than the LP
systems. However, SET and Crystal IS (Green Island, New
York), manufacturers of UV-LEDs, estimate great improve-
ments in the next three to four years. If the projected values
manufacturers are aiming for are met, a UV-LED system
could be a viable and improved option over current LP
systems and fill a gap for low-flow, inexpensive systems in
three to four years (Table 1).
Increasing power output will be necessary for systems to
utilize a reasonable number of LEDs independent of lamp
cost. Each LED requires wiring and other electrical
components such as resistors and heat sinking material.
More LEDs also require a larger system and more materials
that will cost more up front. Maintenance will also be more
difficult with a larger number of LEDs since each device will
need to be monitored to detect failures. This will be
particularly important for systems that require a high flow
rate, where thousands of LEDs may become difficult to
install and maintain. Based on manufacturer expectations,
100 mW (power output) LEDs should be on the market by
2013. Improving the power output based on manufacturer
projections over the next three to four years, shows a large
Figure 5
|
Dose for given flow rate and UV transmittance for E. coli K12 spiked into
PBS and natural water.
Table 1
|
Comparison of current and projected future UV-LEDs with LP systems
UV-LEDs
UV-Tube Sterilight Base case
3– 4 year
projection
MW/Lamp (output) 15,000 10,000 0.36 100
Lifetime (hrs) 9,000 9,000 1,000 10,000
Cost ($/mW)*0.0013 0.0055 664 0.1
Flow rate (mL/min) 6,000 1,890 55.55 55.55
Total mW (output)
†
72.22 72.22
Number of LEDs 201 1
Upfront lamp cost ($) 20 55 47,943 7
3 year cost 60 165 1,260,063 21
20 year cost 389 1,071 8,400,421 123
*
$ values in USD. Current LED cost based on purchase price from SET December 2007.
†
Total power output was calculated by multiplying the flow rate by 1.3 as found in the
prototype testing.
6C. Chatterley and K. Linden
|
UV-LEDs for water disinfection Journal of Water and Health
|
inpress
|
2010
Uncorrected Proof
decrease in the number of LEDs required (from over 200 to
only one LED) for a constantly running household system
that would treat eighty liters per day at 40 mJ/cm
2
with a
UVT as low as 75%.
One of the most desired features of LEDs for disinfec-
tion systems is their long lifetime, particularly for develop-
ing communities, where replacements can be difficult to
come across. According to SET engineers, UV-LEDs in the
germicidal wavelength range currently have very low life-
times of approximately 1,000 hours before 50% power
reduction is reached. Manufacturer projections for the next
three to four years would offer lifetimes equal to that of LP
lamps (10,000 hours by around 2012). However, since they
do not need to warm-up, they can be effectively run
intermittently on demand, increasing the total lifetime ten
to twenty fold assuming no decay due to frequent on-off
switching. LP lamps can be run intermittently as well, but
due to the required warm-up time, water is not available on
demand without the need for storage, a known hygiene risk.
Because UV LEDs were shown to be as effective for
bacterial inactivation as LP UV lamps, the most influential
factor to improve the adoption of UV-LED disinfection is
cost decrease. Based on manufacturer’s three to four year
projections, the cost will decrease over 1,000 fold to $0.1
per mW in 2013. The large decrease in three-year cost for a
household system brings the total cost to $190 USD, which
is almost as cheap as the three year cost for the Sterilight
system lamps at $165 USD (Table 1).
Combining projected improvements to power output,
lifetime, and cost per mW, results in UV-LEDs being a
feasible option and an improvement over LP systems
around the year 2013 (Table 2). If the projections can be
met, it will be possible to develop a household system that
will treat eighty liters per day at 40 mJ/cm
2
(if UVT of
water greater than or equal to 75%) for $7 USD of upfront
lamp cost, compared to $20 to $55 USD for lamps in the
UV-Tube and Sterilight systems, respectively. The cost
savings will increase yearly with slightly higher lifetime
values of 10,000 hours for the LEDs, versus 9,000 hours for
the LP lamps, resulting in lower yearly replacement costs.
The long-term cost savings can be increased further
and the maintenance required to replace burned out lights
Table 2
|
Effect of improving all three parameters; power output, lifetime, and cost
Vary all Base case Case 1 Case 2 Case 3 Case 4 Case 5
MW/lamp 0.36 2 5 10 50 100
Lifetime (hrs) 1,000 2,000 4,000 6,000 8,000 10,000
Cost ($/mW) 664 332 100 10 1 0.1
Total lifetime (days) 42 83 167 250 333 417
Total lifetime (years) 0.11 0.23 0.46 0.68 0.91 1.14
Number of LEDs 201 37 15 8 2 1
Total lamp cost (upfront) USD 47,943 23,975 7,222 722 72 7
Cost for 3 years USD 1,260,063 315,068 47,450 3,163 237 19
Table 3
|
Effect of increasing flow rate for future UV-LED systems
Vary flow rate Constant Case F1 Case F2 Case F3 Case F4 Case F5
Flow rate (mL/min) 55.55 500 1,000 1,890 5,000 6,000
Hours/day 24.0 2.7 1.3 0.7 0.3 0.2
Total lifetime (days) 417 3,750 7,500 14,175 37,500 45,000
Total lifetime (years) 1.14 10.27 20.55 38.84 102.74 123.29
Total mW 72.22 650 1,300 2,457 6,500 7,800
Number of LEDs 1 7 13 25 65 78
Total lamp cost (upfront) 7 65 130 245.7 650 780
7C. Chatterley and K. Linden
|
UV-LEDs for water disinfection Journal of Water and Health
|
inpress
|
2010
Uncorrected Proof
can be decreased, by increasing the system flow rate and
turning on the LEDs intermittently as water is needed
without the need for storage where recontamination can
easily occur (Table 3).
CONCLUSIONS
UV-LEDs are an effective technology to inactivate E. coli in
water, comparable to LP UV mercury vapor lamp technol-
ogy. The efficiency of UV-LEDs at 265 nm was not found to
be statistically different from low-pressure UV sources.
A ten LED prototype served as a proof-of-concept for
flow-through water treatment, but currently UV-LEDs in
the germicidal wavelength range are much too expensive,
low power and have short lifetimes. For UV-LEDs to be a
feasible option for field implementation the cost needs to
decrease and the power output needs to increase substan-
tially. However, according to an analysis of manufacturer
projections, UV-LEDs should be a viable and economic
option within four years, by around 2013. Once UV-LEDs
become a viable option, there are numerous possible
applications for UV-LED technology within the water
sector, including low-pressure lamp replacement in drink-
ing water, wastewater, and gray water treatment systems, in
solar powered or plug-in point-of-use treatment systems in
rural and urban households in developing countries, and in
portable systems.
ACKNOWLEDGEMENTS
This work was funded by a fellowship to Ms. Christie
Chatterley from the National Water Research Institute
(NWRI) and a grant from the University of Colorado
Engineering Excellence Fund (EEF). The authors wish to
thank Dr. Kevin McCabe (University of Colorado), Mr. Tim
Bettles (Crystal IS), Tim May (University of Colorado),
Naomi Levine (Boulder, CO), Christina Barstow (University
of Colorado) and David Sparkman (University of Colorado)
for their assistance in the research and data analyses.
REFERENCES
Bettles, T., Schujman, S., Smart, J. A., Liu, W. & Schowalter, L.
2007 UV Light Emitting Diodes; Their Applications and
Benefits. Proceedings. International Ultraviolet Association
Conference. Los Angeles, CA.
Bolton, J. & Linden, K. 2003 Standardization of methods for
fluence (UV Dose) determination in bench-scale UV
experiments.J. Environ. Eng. ASCE 129(3), 209–215.
Brownell, S. A., Chakrabarti, A. R., Kaser, F. M., Connelly, L. G.,
Peletz, R. L., Reygades, F., Lang, M. J., Kammen, D. & Nelson,
K. L. 2008 Assessment of a low-cost, point-of-use, ultraviolet
water disinfection technology.J. Water Health 6(1), 53 –65.
Crawford, M., Banas, M., Ross, M., Ruby, D., Nelson, J., Boucher,
R. & Allerman, A. 2005 Final LDRD report: Ultraviolet Water
Purification Systems for Rural Environments and Mobile
Applications. Sandia Report SAND2005-7245.
Fewtrell, L. & Colford, J. 2004 Water, Sanitation and Hygiene:
Interventions and Diarrhoea. A Systematic Review and
Meta—analysis. HNP Discussion Paper. The World Bank,
Washington, DC.
Gaska, R. 2007 Deep Ultraviolet Light Emitting Diodes for Water
Monitoring and Disinfection. Proceedings, International
Ultraviolet Association Conference. Los Angeles, CA.
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, 2291–2298.
Linden, K. G., Thurston, J., Schaefer, R. & Malley, J. P., Jr. 2007
Enhanced UV inactivation of adenoviruses under polychromatic
UV lamps.Appl. Environ. Microbiol. 73(23), 7571 –7574.
Mamane-Gravetz, H., Linden, K., Cabaj, A. & Sommer, R. 2005
Spectral sensitivity of Bacillus subtilis spores and MS2
coliphage for validation testing of UV reactors for water
disinfection.Environ. Sci. Technol. 30, 7845–7852.
Pruss, A., Kay, D., Fewtrell, L. & Bartram, J. 2002 Estimating the
burden of disease from water, sanitation, and hygiene at a
global level. Environ. Health Perspect. 110(5), 537–542.
Sinha, R. & Ha¨ der, D. 2002 UV-induced DNA damage and repair: a
review.Photochem. Photobiol. Sci. 1, 225 –236.
Sobsey, M., Stauber, C., Casanova, L., Brown, J. & Elliot, M. 2008
Water filtration: a practical, effective solution for providing
sustained access to safe drinking water in the developing
world.Environ. Sci. Technol. 42, 4261–4267.
Standard Methods for the Examination of Water and Wastewater
1998 20th edition. American Public Health Association/
American Water Works Association/Water Environment
Federation, Washington, DC.
U.S. Environmental Protection Agency (US EPA) 2006 Ultraviolet
disinfection guidance manual. EPA 815-R-06-007, Office of
Water, Washington, DC.
WHO 2002 The World Health Report 2002—Reducing Risks, Promoting
Healthy Life. World Health Organization, Geneva, Switzerland.
First received 26 August 2009; accepted in revised form 2 January 2010. Available online 9 March 2010
8C. Chatterley and K. Linden
|
UV-LEDs for water disinfection Journal of Water and Health
|
inpress
|
2010
Uncorrected Proof
