Document

Transcription

Document
Home
Search
Collections
Journals
About
Contact us
My IOPscience
Spectral irradiance measurement and actinic radiometer calibration for UV water disinfection
This content has been downloaded from IOPscience. Please scroll down to see the full text.
2014 Metrologia 51 S282
(http://iopscience.iop.org/0026-1394/51/6/S282)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 176.9.124.142
This content was downloaded on 03/12/2014 at 14:56
Please note that terms and conditions apply.
Bureau International des Poids et Mesures
Metrologia
Metrologia 51 (2014) S282–S288
doi:10.1088/0026-1394/51/6/S282
Spectral irradiance measurement and
actinic radiometer calibration for UV water
disinfection
Peter Sperfeld1, Bettina Barton1, Sven Pape1, Anna-Lena Towara1,
Jutta Eggers2 and Gabriel Hopfenmüller3
1
Physikalisch-Technische Bundesanstalt, Braunschweig & Berlin, Germany
DVGW-Technologiezentrum Wasser, Karlsruhe, Germany
3
sglux SolGel Technologies GmbH, Berlin, Germany
2
E-mail: peter.sperfeld@ptb.de
Received 6 June 2014, revised 29 July 2014
Accepted for publication 10 September 2014
Published 20 November 2014
Abstract
In a joint project, sglux and PTB investigated and developed methods and equipment to
measure the spectral and weighted irradiance of high-efficiency UV-C emitters used in water
disinfection plants.
A calibration facility was set up to calibrate the microbicidal irradiance responsivity of
actinic radiometers with respect to the weighted spectral irradiance of specially selected lowpressure mercury and medium-pressure mercury UV lamps.
To verify the calibration method and to perform on-site tests, spectral measurements were
carried out directly at water disinfection plants in operation. The weighted microbicidal
irradiance of the plants was calculated and compared to the measurements of various actinic
radiometers.
Keywords: radiometery, UV water disinfection, microbicidal irradiance, actinic radiometer,
UV lamps, UV-C
(Some figures may appear in colour only in the online journal)
1. Introduction
More and more water disinfection systems are equipped with
high-efficiency UV emitters, which have to fulfil specific
requirements for their spectral irradiance. A high irradiance
should be achieved in the spectral range covering the microbicidal action spectrum between 240 nm and 290 nm, whereas
below 240 nm, the irradiance should be low, in order to avoid
any possible photochemical processes leading to by-product
formation. When disinfecting water, a prescribed reduction
equivalent dose (RED) has to be applied to guarantee a sufficient microbicidal effect [1]. Depending on the water flowrate
and transmittance, defined irradiances in the UV-C spectral
range thus have to be achieved and continuously measured.
The irradiance is often measured by means of specially
adapted actinic radiometers. The radiometer response has
to represent the weighted microbicidal irradiance of the UV
0026-1394/14/06S282+7$33.00
emitters. However, the spectral responsivity does not exactly
match the microbicidal action spectrum and the readout of
the radiometer has to be corrected by its spectral mismatch.
The direct spectral responsivity calibration of such actinic
radiometers is very limited due to the high spectral power
needed for such measurements. Thus, a calibration facility
has been constructed, where actinic radiometers are calibrated with respect to the spectral irradiance of typical highpower UV sources.
Actinic radiometers that monitor UV lamps cannot provide
any information on the spectral distribution of the lamps in
the plants. Especially when characterizing new UV devices,
or after replacing a lamp, spectral measurements are, however, of great importance. This is the only way to rule out, for
example, radiation in the spectral range below 240 nm where
undesirable chemical byproducts are generated. Spectral
measurements at high UV irradiances, however, place high
S282
© 2014 BIPM & IOP Publishing Ltd Printed in the UK
P Sperfeld et al
Metrologia 51 (2014) S282
requirements on the spectroradiometers used—and the corresponding entrance optics.
With specially adapted measuring instruments, and in combination with actinic radiometers, a reliable traceability of the
weighted irradiance measurements is possible.
In this article the theoretical background for a source-based
calibration method is presented. This method requires traceable spectral irradiance measurements and reliable standard
sources of a calibration facility to calibrate sensors for high
UV irradiances. In addition, spectral measurements at disinfection plants and the comparison to detector measurements
were carried out to verify the feasibility and practicability of
on-site measurements.
In most cases, however, it is not possible to obtain an
optimum adaptation of the relative spectral responsivity of a
radiometer to the microbicidal action spectrum. Often, SiC
photodiodes are used, whose spectral responsivity for longer
wavelengths is cut off by means of a suitable interference filter.
This, however, often reproduces the curve of the microbicidal
action spectrum to a limited extent only. The deviations of
the spectral slopes from the spectral responsivity and from the
actinic action spectrum have different effects depending on
the radiation sources used. This can be expressed by means of
the spectral mismatch factor mS:
2. Calibration methods
mS =
∫ smic,rel (λ) ⋅ Eλ,S,rel (λ) dλ
.
∫ Amic (λ) ⋅ Eλ,S,rel (λ) dλ
(4)
(1)
Here, ES,rel(λ) is the relative spectral irradiance of the radiation
source S and Eλ,S(λ) = ES,0·Eλ,S,rel(λ).
Hence, if the spectral mismatch factor mS of a radiometer
for a defined radiation source S and the actinic irradiance of
the source are known, the corrected responsivity of the radiometer can be determined:
smic,S = s0 ⋅ mS .
(5)
An actinic radiometer has a specially adapted spectral
responsivity smic(λ) whose relative spectral curve smic,rel(λ) is,
ideally, identical to the microbicidal function Amic(λ). When
irradiated with the radiation source S, the following photocurrent is obtained for the radiometer:
If the radiation source S is a low-pressure mercury lamp,
then mS = 1, since the spectrum of the source S basically only
exists at the wavelength λ0.
For all other sources, mS or at least smic,S has to be determined separately.
The microbicidal effect of UV radiation is defined by a corresponding action spectrum Amic(λ) which corresponds to the
inactivation of the spores of bacillus subtilis [1]. In the case of
a radiation source S of known spectral irradiance Eλ,S(λ), the
microbicidal irradiance is defined as
Emic,S =
∫ Amic (λ) ⋅ Eλ,S (λ) dλ .
∫
imic , S = s0 ⋅ smic ,rel (λ ) ⋅ Eλ, S (λ ) dλ
(2)
where s0 is the absolute value of the irradiance responsivity
of the radiometer, measured in (A W−1 m−2). When calibrating
and using the radiometer, the responsivity smic,abs is interesting with regard to the microbicidal irradiance Emic which
is defined as
smic,abs,S =
imic,S
= s0 ⋅
Emic , S
∫ smic ,rel (λ) ⋅ Eλ,S (λ) dλ
.
∫ Amic (λ) ⋅ Eλ,S (λ) dλ
(3)
In the case of an ideal actinic radiometer, the following
applies: smic,rel(λ) = Amic(λ), and thus smic,abs,S = s0. This means
that the absolute value of the radiometer responsivity only
needs to be calibrated against a single wavelength in order
to determine the responsivity with regard to the microbicidal
irradiance for any random radiation emitter.
This type of calibration is often carried out with low-pressure mercury lamps which, in the UV-C spectral range, mostly
exhibit only one defined spectral line at λ0 = 253.6 nm. If only
low-pressure mercury lamps are used for water disinfection,
then this conventional calibration is fully sufficient. For this
purpose, low-pressure lamps of known irradiance can be used
for source-based calibration. Or in the case of detector-based
calibration, a radiometer of known responsivity is used and
compared under irradiation with a low-pressure lamp by
directly substituting it for the radiometer to be calibrated.
3. Spectroradiometric measurements
Often, the spectral responsivity smic(λ) of a radiometer cannot
be determined directly since the sensors are relatively insensitive to irradiation in order to be able to measure high irradiances at all. In the case of a spectral calibration, however,
the spectral irradiances involved are often comparatively low,
which generate a signal with a low signal-to-noise ratio in
radiometers.
In many cases, actinic radiometers are therefore calibrated directly with a radiation source of known microbicidal
irradiance.
For this type of source-based calibration, it is necessary to
know the exact spectral irradiance Eλ,S(λ). Then, the radiometer’s responsivity to microbicidal irradiance for the radiation
source S can be determined by measurement:
smic,abs, S =
imic, S
∫ Amic (λ) ⋅ Eλ,S (λ) dλ
.
(6)
Hereby, imic,S is the radiometer’s signal or indication at the
microbicidal irradiance according to equation (1) which is
formed by the integral in the denominator.
Traceable spectral measurements of high-power UV
sources are rather challenging. The spectral irradiance of
such lamps is higher by several orders of magnitude than the
spectral irradiance of typical transfer standard lamps. Typical
spectral irradiances of UV emitters used in water disinfection
S283
P Sperfeld et al
Metrologia 51 (2014) S282
Lam p:
10
: H a lo g e n la m p
sp e c tra l irra d ia n ce / W /(m ²· n m )
: D e u te riu m la m p
: H g m e d iu m p re s s u re
1
: H g lo w p re s s u re
0 ,1
0 ,0 1
1 0 -3
1 0 -4
1 0 -5
1 0 -6
1 0 -7
300
400
500
w a v e le n g th / n m
600
700
800
Figure 1. Spectral irradiance of standard lamps compared to high-power UV lamps measured using the array spectroradiometer described
in this paper.
plants are shown in figure 1 compared to the spectral irradiance distribution of standard lamps used in spectroradiometry.
For spectral measurements on such different radiation
sources, high requirements must be placed on the spectroradiometers used and on their entrance optics:
•The instruments must exhibit high dynamics and linearity
up to four orders of magnitude of the signal.
•They should exhibit sufficient repeatability and reproducibility and should be largely insensitive to variations of
the ambient conditions (temperature, humidity).
•The entrance optics and the sensors must have a high
endurance up to the very high UV irradiances.
•Sufficient stray light suppression with regard to longwave radiation is indispensable.
•The entrance optics must have a defined reference plane,
and the field of view should correspond to that of the
respective actinic radiometers.
In order to perform measurements directly on site at water
disinfection plants, the spectrometer must be easily transportable and mobile calibration possibilities should be available.
For this type of utilization, two spectroradiometers have
been modified and characterized at PTB. The first instrument
is a fast-scanning double monochromator system with fibrecoupled entrance optics. This device has a high stray light
suppression and, by combining a photomultiplier with neutral
density filters, it exhibits a high dynamic range. To determine
the microbicidal irradiance, the lamp spectrum is measured in
the wavelength range from 200 nm to 360 nm. For these measurements, the spectrometer is adjusted to a spectral bandwidth
of 1 nm. Recording a spectrum in wavelength steps of 0.5 nm
takes approximately 90 s, regardless of the irradiance level.
The second spectrometer used is a charge-coupled device
(CCD) array spectroradiometer with a spectral range from
200 nm to 800 nm and a spectral bandwidth of 2.4 nm. This
device is also equipped with fibre-coupled entrance optics.
The integration time of the measurements varies between
15 ms and 10 s, depending on how high the maximum irradiance is. The device turns out to be very linear with regard
to the integration time variation. As for most array spectroradiometers, however, this device has a rather low stray light
suppression, with effects especially seen in the UV spectral
range. E.g. when incandescent lamps are used as calibration
standards, a stray light level clearly higher than 20% may
occur below 300 nm. The stray light behaviour of the array
spectroradiometer used was therefore thoroughly characterized and considerably reduced with the aid of a stray light
correction matrix and an out-of-range correction [2].
An integration time variation is insufficient to bridge the high
difference in dynamics between the calibration sources and the
UV lamps to be characterized. Hence, neutral density filters
have been integrated into both spectrometers. These allow an
attenuation of the induced radiation by up to 3 orders of magnitude. Neutral density filters, however, are optimized mainly
for the spectral range above 300 nm where they really exhibit
a largely spectrally independent transmission curve. In the
UV-C spectral range, these filters show a strongly decreasing
transmission by up to −2% nm−1. The spectral transmission of
the neutral density filters used was therefore measured several
times in the spectral range from 200 nm to 360 nm before and
after operation under high UV irradiation. Figure 2 shows the
transmission of the three density filters used in the array spectroradiometer with the optical densities OD1–OD3. It was also
demonstrated that the transmission of the filters remains constant over time, in spite of intermittently high UV irradiances.
The two spectroradiometers have identical entrance optics
for the spectral measurements. A flat quartz glass diffuser
is mounted into a housing for an actinic radiometer (figure
3), as used in water disinfection plants [3]. This allows the
entrance optics of the spectroradiometers to be inserted
S284
P Sperfeld et al
Metrologia 51 (2014) S282
: OD1
: OD2
: OD3
0,1
transmission
0,05
0,01
0,005
10-3
5·10-4
200
300
400
500
600
wavelength / nm
700
800
Figure 2. Transmission of several neutral density filter types incorporated into the two spectroradiometers used.
Figure 3. Schematical drawing of the adapted spectrometer
entrance optics. (1) housing, (2) diffuser plate, (3) optical fibre.
directly on site. Hereby, the 160° field of view of the diffusers corresponds to that of the radiometers. The diffusers
were thoroughly characterized and tested, especially as to
their UV resistance [4]. Quartz glass-fibre bundles, which
are located directly behind the diffusers, lead the incident
radiation onto the entrance slit of the spectroradiometers.
The radiation incident at the entrance slit has already been
attenuated by the diffusers to such an extent that it is proven
that it has no harmful effect on the optical elements of the
spectroradiometers.
Thus, two clearly different spectroradiometers are available to carry out spectral measurements. After application of
all corrections and transmission adaptations within the scope
of their measurement capabilities, these spectroradiometers
should provide measurement results that are in good agreement with each other for different lamps.
4. Calibration facility
It has turned out that due to the spectral mismatch in commercially available actinic radiometers, the calibration factors must be determined separately, at least for each type of
lamp used [5]. Therefore, commercially available UV lamps
designed specifically for UV disinfection have been characterized and their suitability as a calibration standard has been
investigated [6]. To calibrate sensors for UV water disinfection plants based on medium-pressure mercury emitters as
well as on low-pressure mercury sources, a 1000 W mediumpressure lamp and a 40 W low-pressure lamp were selected.
Figure 4. The transfer standard for high-power UV irradiance. A
medium pressure lamp (1) is mounted at the top, the low pressure
mercury lamp (2) in the middle can be removed. The sensor mount
with baffle tube (3) is located at the bottom of the chamber.
The medium-pressure mercury lamp is mounted on top
of a ventilated cabinet and the low-pressure source can be
mounted underneath (figure 4). The measurement chamber
is double-walled and has reflecting internal walls. A fan
blows air through the gap between the double wall to cool
the internal chamber without causing a draught inside. After
the working temperature has been reached, this uniform background ventilation provides stable temperature conditions
inside the chamber. When using the 1000 W lamp, temperatures between 58 °C at the bottom and 105 °C at the top of the
internal chamber are reached.
Both lamps were characterized with regard to their stability
and reproducibility, and their temperature curve was measured
S285
P Sperfeld et al
Metrologia 51 (2014) S282
The spectral irradiance Eλ,S(λ) of the two types of lamps
used, as presented in figure 1, was measured repeatedly with
the two adapted spectroradiometers, and the microbicidal irradiance Emik,S was calculated. The low-pressure mercury lamp
achieves approximately Emik,Hg-LP = (3.3 ± 0.1) W m−2,
whereas the medium-pressure lamp reaches Emik,Hg-MP =
(45 ± 2) W m−2. Hereby, the microbicidal irradiances, which
were calculated independently of each other with the two
spectroradiometers, deviate from each other by less than 1%.
Thus at this measuring facility, irradiances are attained
which allow UV radiometers for water disinfection to be calibrated directly. In industrial water disinfection plants, however, much higher irradiances may occur.
5. On-site spectral measurements
Figure 5. Schematic drawing of the sensor mount (1) and the baffle
tube (2). Monitor photodiodes are located right and left of the centre
(3). At the top of the baffle tube, a long-pass filter can be placed (4).
A mechanical shutter (5) can be rotated to completely block the UV
radiation.
at different locations inside the chamber. The electrical power
of the 1000 W medium-pressure lamp is regulated by its mains
supply. UV monitor sensors register a stable signal after temperature balance has been reached inside the chamber. This
process takes approximately 90 min. Upon restart, the reproducibility of the irradiance is, after another pre-warming
period, better than 0.5%. After this, a stability of the irradiance
of better than 0.2% is attained during the operation of the lamp.
The low-pressure mercury lamp, which is operated with
DC voltage, reaches stable operating conditions after approximately 120 min. Only then has the temperature balance inside
the measuring chamber been reached at which a temperature
of approximately 45 °C at the lamp is attained. The reproducibility of the microbicidal irradiance is then also better than
0.5% here, and the stability lies at around 0.2%.
A holder is located at the bottom of the chamber. This holder
can accommodate both UV sensors and the entrance optics of
the spectroradiometers (figure 5). Towards the top, it is closed
by means of a baffle tube which limits the sensors’ field of
view. On the upper side of the baffle tube, there is a fixture for
a filter and a mechanical shutter for dark measurements. Two
UV monitor sensors monitor the incident radiation. One of the
sensors is permanently exposed to the radiation; the other one
is located behind the shutter and carries out measurements at
the same time as the measurements with the UV radiometers
or spectrometers.
The baffle tube limits the field of view in such a way that
only the direct irradiation of the lamps is incident on the UV
radiometers and reflections from the walls or other components are ruled out. Thus, identical optical conditions are provided for all sensors and entrance optics, although the total
field of view of the sensors is not used at calibration.
A long-pass filter with 235 nm cut-on wavelength is inserted
into the filter holder at the top end of the baffle tube. This fused
silica filter with a steep transmission change between 230 nm
and 250 nm blocks most of the radiation below 240 nm.
In Germany, water disinfection plants undergo thorough tests
before they are granted approval. TZW, the Water Technology
Centre (Technologiezentrum Wasser) of DVGW, the German
Technical and Scientific Association for Gas and Water
(Deutscher Verein des Gas- und Wasserfaches e.V.) carries
out the inspection of new plants. For this purpose, TZW operates a testing laboratory for UV devices at the water works of
St.-Augustin-Meindorf, where large-scale UV plants can be
tested at flowrates of up to 3500 m3 h−1. By means of microbiological test procedures, the microbicidal effectiveness of
the devices can be tested under various operating conditions.
Hereby, the German Drinking Water Ordinance (TrinkwV
2001, Deutsche Trinkwasserverordnung) requires that water
flowing through such UV plants is subjected to a RED of
400 J m−2. For this reason, numerous correspondingly strong
UV lamps must be available at devices operated at a high
flowrate. The irradiance of each lamp must be monitored by
means of UV radiometers. The lamps’ power can be adjusted
as a function of the flowrate. TZW checks the adjusting
mechanisms of new devices and the irradiances of the UV
lamps. The measurements are performed with calibrated UV
radiometers. Direct spectral measurements were previously
not feasible at large-scale UV plants.
With the adapted spectroradiometers, it was possible to
test directly at TZW’s test rig whether spectral measurements
are feasible at all [7]. For this purpose, a mobile measuring
assembly of PTB (consisting of the spectroradiometers, several UV radiometers and calibration sources) was set up at
the water works. There, with temperatures around 12 °C,
the ambient conditions were clearly different from those
at PTB’s laboratory. In addition, the geometric conditions
at a UV plant and, for instance, the effect of flowing water
cannot be simulated at the radiometry lab. Direct comparison
with reference radiometers of TZW is therefore an important indicator of the reliability of the field measurements.
Microbicidal irradiances were determined at several devices
and compared with each other.
First measurements at a small UV reactor with low-pressure mercury lamps, a flowrate of 6 m3 h−1 and a microbicidal
irradiance of 63 W m−2 ± 2 W m−2 showed a good agreement between the spectroradiometers, the UV radiometers
S286
P Sperfeld et al
Metrologia 51 (2014) S282
Table 1. Measurement results of the microbicidal irradiance at a UV reactor with 48 kW medium-pressure mercury lamps, at different
power settings.
Relative
electrical power
TZW sensor
PTB Spectrorad. 1
PTB Spectrorad. 2
PTB Sensor 2
PTB Sensor 3
PTB Sensor 4
30%
50%
100%
181 W m−2
350 W m−2
827 W m−2
206 W m−2
401 W m−2
969 W m−2
193 W m−2
378 W m−2
906 W m−2
200 W m−2
386 W m−2
910 W m−2
196 W m−2
379 W m−2
898 W m−2
203 W m−2
389 W m−2
920 W m−2
rel. electrical pow er
: 30 %
0,1
: 50 %
: 100 %
E ( λ ) / E m ik
0,01
10 -3
10 -4
10 -5
200
300
400
500
W avelength / nm
600
700
800
Figure 6. Normalized spectral irradiance of a UV reactor with 48 kW medium-pressure mercury lamps at different power settings.
and TZW’s reference radiometers. The standard deviation of
all measurement results was below 2.5%. The result at this
reactor was verified also at 50% of the irradiance power and a
flowrate of 0.6 m3 h−1.
Larger water disinfection plants are increasingly being
equipped with medium-pressure mercury lamps. Their compact
design allow a considerably higher energy density in minimum
space. During the joint measurements at TZW’s test rig, we
had the opportunity to use a UV reactor with 48 kW mediumpressure lamps in total. The device’s electrical power was set
between 30% and 100%, so that microbicidal irradiances from
approximately 200 W m−2 to 900 W m−2 were reached at the
measurement points. These irradiances lie much higher than the
values attained at the calibration facility in the lab.
Table 1 shows that mainly consistent agreement was achieved
between all the measuring instruments. The standard deviation
of all PTB measurement results amounts to less than 3%.
TZW’s sensor shows an indication of the microbicidal irradiance which is lower by approximately 10%. This is due to
the fact that the sensor was calibrated by the manufacturer
with respect to a low-pressure mercury lamp and the mismatch
to the spectrum of a medium-pressure mercury lamp was not
taken into account in these measurements. For this type of
sensor, this deviation amount is expected and usually leads
to an indication of the microbicidal irradiance which is too
low. The fast-scanning spectroradiometer (PTB Spectrorad.
1) shows higher values than the mean value of the PTB
measurements. The deviation increases slightly with rising
irradiance. Both the clearly different ambient conditions and
a possible non-linearity of the photomultiplier at these high
irradiances can be responsible for this behaviour.
In addition to the measured microbicidal irradiances, the
spectral curves of the irradiance provide important information concerning the tested UV plants. Figure 6 shows
the spectral slope of the measurements at different power
settings of the device, normalized to the resulting microbicidal irradiance. Here, it can be seen clearly that the variation of the electrical power has nearly no influence on the
relative spectral slope. This allows us to rule out a change
in the microbicidal effectiveness due to a shift of the lamp
spectrum.
This observation justifies the use of UV radiometers to
check the UV plants, as far as the correction factors are correspondingly adapted to medium-pressure lamps. A change
in the electrical power of the lamps does not cause a shift
of this adjustment factor. In addition, figure 6 shows clearly
that the required absorption of the UV radiation is reached
below 240 nm.
It has thus been shown that traceable spectral measurements can be carried out, even at high UV irradiances
at these reactors. This allows, for instance, new types of
radiation sources to be characterized or the change in the
spectral irradiance as a function of the water properties to
be investigated. Hereby, these results have confirmed that
S287
P Sperfeld et al
Metrologia 51 (2014) S282
different sources require different adapted adjustment factors for the UV radiometers. The commercially available
calibration station can hereby be used by testing laboratories as well as by sensor manufacturers to calibrate the UV
radiometers.
Acknowledgments
6. Conclusion
References
For the first time, PTB has carried out traceable spectral
measurements directly on UV disinfection plants in operation. Water disinfection systems using two types of mercurybased UV emitters were measured at the TZW test centre in
St. Augustin, Germany. The resulting calculated microbicidal
irradiances agreed with the measurement results of the reference radiometers used at the test centre.
Additional measurements using reference radiometers calibrated at a newly designed calibration facility confirmed the
introduced source-based calibration method.
Hereby, these results have confirmed that different radiation sources require different adapted adjustment factors for
the UV radiometers. The commercially available calibration
station can hereby be used by testing laboratories as well as by
sensor manufacturers to calibrate the UV radiometers.
In addition, the direct spectral measurements allow new
water disinfection UV plants and new types of lamps to be
investigated and developed.
Combined with newly developed UV transfer standard
lamps, it will be possible to calibrate broadband actinic UV
radiometers in terms of microbicidal irradiance responsivity
for these new types of radiation sources.
[1] Arbeitsblatt W 2006 294-3: UV-Geräte zur Desinfektion in der
Wasserversorgung; Teil 3: Messfenster und Sensoren zur
radiometrischen Überwachung von UV-Desinfektionsgeräten—
Anforderungen, Prüfung und Kalibrierung, Technische
Regel der DVGW, Ausgabe 2006-06, Wirtschafts- und
Verlagsgesellschaft Gas und Wasser mbH, Bonn, June 2006
[2] Sperfeld P, Nevas S and Pape S 2014 Application of a
compact array spectroradiometer for the UV spectral range
Poster presented at NEWRAD 2014 (Helsinki, Finland,
24–27 June 2014) pp 263–4
[3] Hopfenmüller G et al 2013 PTB traceable calibrated reference
UV radiometer for measurements at high irradiance medium
pressure mercury discharge lamps Proc. 2013 EMEA
Regional Conf. (Karlsruhe, Germany, 4–5 June 2013)
[4] Barton B et al 2011 Characterization of new optical diffusers
used in high irradiance UV radiometers Poster Presented at
NEWRAD 2011 Maui, Hawaii, 19–23 September 2011)
[5] Larason T and Ohno Y 2006 Calibration and characterization
of UV sensors for water disinfection Metrologia 43 S151–6
[6] Barton B et al 2013 Developing and setting up a calibration
facility for UV sensors at high irradiance rates Proc. 2013
EMEA Regional Conf. (Karlsruhe, Germany, 4–5 June 2013)
[7] Sperfeld P, Barton B, Pape S, Eggers J and Hopfenmüller G
2013 Traceable measurements of the spectral irradiance of
UV water disinfection plants Proc. 2013 EMEA Regional
Conf. (Karlsruhe, Germany, 4–5 June 2013) pp 122–3
This work was supported in part by the Federal Ministry for
Economic Affairs and Energy based on a decision of the German Parliament under grant no. KF2303704RR9.
S288