METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
METHODS OF CHARACTERIZATION AND
CALIBRATION OF BROAD-BAND UV RADIOMETERS
CIE TC2-47
(For discussion only)
Draft Version :
Date
:
Prepared by
:
3.1
Armin Sperling (Germany)
Gan Xu (PSB) until 2006
Contact:
Tel: (+49) 531 592 4121
Fax: (+49) 531 592 4105
e-mail: armin.sperling@ptb.de
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Table of Content
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
1. Introduction
1.1. Objectives
1.2. Scope
2. Definitions (General)
2.1. Radiometer
2.2. Broadband UV Radiometer
2.3. Detector
2.4. Spectral distribution
2.5. Action spectrum
2.6. Effective irradiance, radiant power and radiant exposure
2.7. Spectral responsivity
2.8. Effective radiant/irradiance responsivity
2.9. Calibration source
3. Quality indices
3.1. Calibration uncertainty
3.2. Spectral match quality index f1’
3.3. Short-wavelength range response quality index fu
3.4. Long-wavelength range response quality index fr
3.5. Directional response quality index f2
3.6. Linearity quality index f3
3.7. Display unit quality index f4
3.8. Fatigue quality index f5
3.9. Temperature dependence quality index f6
3.10. Modulated radiation quality index f7
3.11. Polarisation response quality index f8
3.12. Non uniform radiation Spatial response quality index f9
3.13. Range change quality index f10
3.14. Total quality index ftot
3.15. Quality class
3.16. Class limits
4. Properties of UV-radiometers
4.1. Effective radiant power/irradiance responsivity
4.2. Spectral matching
4.2.1. Spectral mismatch correction factor
4.2.2. Spectral match characteristic
4.3. Short and long wavelength range response characteristic fu and fr
4.4. Directional response
4.4.1. General
4.4.2. Measurement
4.4.3. Characterisation for spherical irradiance E
4.4.4. Characterisation for spherical irradiance E0
4.5. Linearity
4.5.1. General
4.5.2. Measurement
4.5.3. Characterisation
4.6. Display unit
4.6.1. General
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
4.6.2. Measurement
4.6.3. Characterisation
4.7. Fatigue
4.7.1. General
4.7.2. Measurement
4.7.3. Characterisation
4.8. Temperature
4.8.1. General
4.8.2. Measurement
4.8.3. Characterisation
4.9. Modulated radiation
4.9.1. General
4.9.2. Measurement
4.9.3. Characterisation
4.10. Polarisation
4.10.1. General
4.10.2. Measurement
4.10.3. Characterisation
4.11. Non uniform radiation
4.11.1. General
4.11.2. Measurement
4.11.3. Characterisation
4.12. Range response
4.12.1. General
4.12.2. Measurement
4.12.3. Characterisation
4.13. Total quality index
4.13.1. General
4.13.2. Characterisation
5. Specification of broadband UV radiometers
5.1. Content of data sheets
5.2. Defining a class for a radiometer
5.3. Class limits for irradiance meters
6. Calibration of broadband UV radiometers
6.1. Measurement of spectral responsivity
6.2. Measurement of spectral irradiance and spectral distribution
6.3. Calibration of effective responsivity
6.4. Calibration standards
6.4.1. Source standards
6.4.2. Detector standards
6.4.2.1.Transfer/working standards
6.5. Calibration conditions
6.6. Calibration interval
6.7. Uncertainty evaluation
6.7.1. Method
6.7.2. Sources of uncertainty
7. Selection and correct usage of UV radiometers
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
7.1. End use and target action spectrum
7.2. Relative spectral responsivity and spectral mismatch
7.3. Source type and spectral distribution
7.4. Spectral correction factor
8. References
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
1. Introduction
1.1. Objectives
Broadband UV radiometers (also called integral measuring UV radiometers) are widely
used for measuring UV irradiance or radiant exposure in various industrial, medical,
environmental, health and other applications. The accuracy or uncertainty of measurement
achievable using this type of radiometers is influenced by many factors such as the
characteristics of the radiometer, operating conditions, environmental conditions and the
UV sources to be measured. Due to the lack of spectral resolving power, significant
measurement errors may occur without notice if the radiometer used has poor quality
(characteristics) and/or is not correctly calibrated or used. This appeared to be the case
frequently in the past twenty years and problems so caused have raised serious concern.
The purpose of this document is to

Give clear definitions for the individual characteristics which can be used for
evaluating the performance of broadband UV radiometers;

Recommend methods for the characterisation of these meters;

Recommend methods for the calibration of these meters;

Provide a guideline for the selection and correct usage of these meters
This document may be used as a guide for

manufacturers in developing good quality UV meters;

calibration labs in performing calibration of these meters in a standardised and
comparable way;

users in the appropriate usage of these meters to get correct and consistent
measurement results for their applications.
1.2. Scope
The UV radiometers considered in this document are restricted to those working in the
wavelength range from 250 nm to 400 nm, i.e., in the spectral bands of UVA (315 nm to
400 nm), UVB (280 nm to 315 nm) and part of UVC (250 nm to 280 nm). The radiation to
be measured is also restricted to that from artificial UV sources only. Radiometers used for
solar UV measurement are excluded since many additional requirements are important due
to the special features of the source (the sun) and changing environmental, weather and
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
geometrical conditions often encountered during measurements. There are already other
parties such as the World Meteorological Organisation dealing with this issue.
Note1: Although some broadband UV meters work beyond the artificial ranges
defined above, they can be treated in the same way as discussed in this document.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
2. Definitions (General)
The following definitions are used in this document.
2.1. Radiometer
A radiometer is an instrument designed for measuring radiometric quantities such as
irradiance, radiance, radiant exposure or radiant power (flux) in a certain spectral range. It
usually consists of a radiometer head, an electronic signal converter and an output
(display) device. This document refers to radiometers for the measurement of irradiance
and radiant exposure.
2.2. Broadband UV radiometer
A broadband UV radiometer is an instrument for measuring the total (spectrally weighted
and integrated) radiation of a specified solid angle, either irradiance or dose (radiant
exposure, time-integrated) over a limited spectral band in the ultraviolet region from a
radiation source. In this document any radiometer with a FWHM more than 20 nm is
considered to be broadband. Hereafter, a broadband UV radiometer is also called a "UV
meter" in this document.
Note 2: A broadband UV radiometer is used to measure radiation either from a
broadband UV source such as a high-pressure Hg arc lamp or a Xe arc lamp, or
from a line source such as a “black light” bulb, whose spectrum is concentrated
essentially in a single wavelength at about 365 nm or UV-LEDs.
2.3. Radiometer head
That part of a radiometer containing detector(s) and other components such as filter,
diffuser, aperture and attenuator for spectral and spatial corrections of the detector. It may
also include electronic signal converter and amplifier. A radiometer head is also called
"meter head" in this document.
2.4. Detector (of optical radiation)
A detector is a device which generates a measurable physical effect when irradiated by
optical radiation.
2.5. Spectral distribution
The Spectral distribution X    of a radiant quantity X (irradiance, radiant exposure)
over a range of wavelength is the function of the spectral concentration of the radiant
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
quantity over the wavelength as defined in the ILV [1]. It represents the absolute spectral
values of the emitted radiation from a source.
The relative spectral distribution S   is the ratio of the spectral distribution X    of a
radiant quantity X e   of a source to a fixed reference value X 0 of this distribution as
defined in the ILV [1]. Throughout this document, the index “e” will be omitted as a
confusion between photometric and radiometric values will not be possible.
2.6. Action spectrum
An action spectrum Aact   refers to the wavelength dependence of the effectiveness of an
optical radiation in a special photochemical, photo-biological or photo-voltaic
phenomenon (actinic effects). It is essentially a spectral weighting function with a peak
value of unity.
Note 3: The action spectrum does not need to describe a real actinic effect, but an
artificial one, as for instance a square (rectangular) function over a distinct UV
range (UV-A, UV-B or UV-C).
2.7. Effective irradiance, radiant power and effective radiant exposure
Effective irradiance or radiant power generated by a source refers to irradiance/radiant
power spectrally weighted by the action spectrum of the phenomenon of concern. The
effective radiant quantity Xact of a source in terms of an action spectrum Aact   can be
expressed as

X act  K   X     Aact    d ,
(2.1)
0
where X  ( ) is the spectral concentration of the radiometric quantity ( E () for spectral
irradiance and   () for spectral radiant power respectively) produced by the source,
Aact   is the spectral weighting function of the phenomenon of concern (action spectrum).
K is the actinic efficacy, which is normally set to unity. Effective radiant exposure (dose)
is the time integration of the effective irradiance over the period of exposure time.
t2
H act   Eact t   dt ,
(2.2)
t1
2.8. Spectral responsivity
Quotient of the output fraction dY   (e.g. electrical current or voltage) of a detector or
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
radiometer head to the fraction dX   of the radiometric quantity X (depending on the
quantity considered) of an incident monochromatic optical radiation at the wavelength .
s  
dY  
dX  
(2.3)
It is a function of wavelength and may also depend on the spectral irradiance level. The
relative spectral responsivity srel() is the ratio of the spectral responsivity s  to a
given reference value, e.g. its peak value s0 , at the peak wavelength 0 and
consequentially:
s   s0  srel   .
(2.4)
where s() is normalized to 0.
Note 1: A broadband UV radiometer is usually designed to measure effective
irradiance and/or effective radiant exposure according to a specific action spectrum.
This requires the relative spectral responsivity of its meter head to match the action
spectrum of concern, i.e. srel() = Aact(..
2.9. Effective radiant power/irradiance responsivity
The output signal of a UV-radiometer is given by:

Y   X    s d
(2.5)
0
and therefore the effective responsivity of the meter with respect to a specific radiant
quantity is the ratio of the output of the meter head to the effective radiant quantity
(equation. 2.1):

sact
s 
Y

 0  0
X act K
X    srel  d
 X     A  d
act
0
(2.6)
That is, the calibration of a UV irradiance meter requires the determination of its effective
irradiance responsivity.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Note 2: From equation (2.6) follows that the effective responsivity strongly depends
on the relative spectral distribution of the source used to produce the radiometric
quantity X, if the relative spectral responsivity of the broadband UV radiometer
being used is not exactly equal to the relative distribution of actinic action spectrum.
This has especially to be considered, if the calibration source differs from the
sources where the UV radiometer is used for measurements.
2.10. Calibration Source
This refers to the radiation source used to produce the required radiant quantity (irradiance,
radiance or radiant power) at the meter head during calibration.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
3.
Quality indices
The following quality indices for broadband radiometers are used to characterize the
accuracy and uncertainty of measurements and performance of radiometers. The name of
each quality index has been taken from the physical effect that influences its value for easy
understanding of its meaning. The names, their symbols and abbreviations are kept
consistent with those used for the characterisation of photometers in CIE Publication
No.69 whenever possible.
3.1. Calibration uncertainty ucal
The calibration uncertainty ucal describes the uncertainty provided by the calibration of the
radiometer. It must include all uncertainty components of all quantities used for the
calibration. It should be the expanded uncertainty evaluated at a confidence level
approximately 95% with a proper coverage factor.
3.2. Spectral mismatch index f1’
The spectral mismatch index f1’ describes the deviation of the relative spectral responsivity
of the radiometer head from the designated action spectrum.
3.3. Short-wavelength range response index fu
The short-wavelength range response characteristic fu describes the responsivity of the
radiometer head at wavelengths shorter than the spectral range of the designated action
spectrum.
3.4. Long-wavelength range response index fr
The long-wavelength range response characteristic fr describes the responsivity of the
radiometer head at wavelengths longer than the spectral range of the designated action
spectrum.
3.5. Directional response index f2
The directional response characteristic f2 describes the responsivity of the radiometer head
to radiation incident at angles other than perpendicular (cosine law).
3.6. Linearity index f3
The linearity characteristic f3 describes the variation of the responsivity of the radiometer
for different levels of irradiance.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
3.7. Display unit index f4
The display unit characteristic f4 describes the influence of the display unit of the
radiometer on the measurement results.
3.8. Fatigue index f5
The fatigue characteristic f5 describes the reversible change of the responsivity of the
radiometer over time under constant irradiation.
3.9. Temperature dependence index f6
The temperature dependence characteristic f6 describes the influence of ambient
temperature that is different from calibration temperature.
3.10. Modulated radiation index f7
The modulated radiation characteristic f7 describes the influence of modulated radiation at
various frequencies, as opposed to that of DC-operated lamps.
3.11. Polarisation response index f8
The polarisation response characteristic f8 describes the influence of polarised radiation on
the responsivity of the radiometer.
3.12. Spatial response index f9
The spatial response characteristic f9 describes the influence of non-uniform irradiation of
radiometers.
3.13. Range change index f11
The range change characteristic f11 describes the influence of range settings of display units
or amplifiers.
3.14. Total quality index ftot
The total quality index ftot is a general measure for a broadband radiometer
3.15. Quality class
A measure for the quality of a UV radiometer. A radiometer belongs to that class where all its
indices do not exceed the given class limits.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
3.16. Class limits
Maximum values for quality indices for each class. The class limits are given in a table for
irradiance meters.
NOTE: Class limits must not be understood as “maximal errors” for the corresponding
indices of the radiometer. Generally the actual measurement uncertainty can be smaller,
but sometimes also can be larger for certain measurement tasks
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
4. Properties of UV-radiometers
4.1. Effective radiant power/irradiance responsivity against a particular source
In equation 2.6 it was defined that the effective responsivity sact of a UV meter in terms of
a particular source is determined by its spectral responsivity s() , the action spectrum
Aact() of the considered effect and the spectral density distribution X    of the radiant
quantity of the source. This equation can be rewritten to

sact 
 X     s d
0

K   X     Aact  d

s 
 0  0
K
0
S    srel  d
 S    A  d
(4.1)
act
0
where
S   relative spectral density distribution of the radiant quantity X of the source,
s0, s rel()
absolute value and relative spectral responsivity distribution of the meter
head.
This points out that since X    appears in both numerator and denominator of equation
2.5, only the relative spectral distribution S() of the source is required for the
determination of sact .
Correspondingly, the effective irradiance responsivity s E is given by
sact  sE , act 
Y
Y  A  cos 

Eact
act
(4.2)
where
A
receiving area of the radiometer head
Y
output of the meter head (or meter) when overfilled by the UV radiation
Eact
effective actinic irradiance at the reference plane of the meter head
act
effective actinic radiant flux at the reference plane of the meter head

angle between the surface normal of the detector and the direction of
irradiation
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Therefore the calibration of UV meters require the accurate determination of its sact for a
specified source. Many existing radiometers have only poor cosine correction and hence
their angle dependent response for non-perpendicular incidence may vary significantly
from that given by equation (4.2).
Ideally, the relative responsivity of a UV meter should match that of the desired action
spectrum. However, equation (4.1) implies that even a meter with a relative response
which totally mismatches the action spectrum could still be calibrated to measure a
particular source if that source has significant output in the spectral range of the meter’s
spectral response. In this case, the calibration source and the particular source being
measured must have the same relative spectral distribution to get reliable results. However,
such a meter could be completely unsuitable to monitor e.g. the aging of even this
particular source, because the relative spectral distribution of the source might change
wavelength-dependent during its lifetime.
Note 3: For a given action spectrum, the effective responsivity of a meter is in
general not a constant since its value still depends on the spectral distribution of the
source unless it’s normalised relative spectral responsivity defined by equation (2.6)
is identical to the action spectrum considered, which is generally very difficult to
achieve. Therefore the calibration and determination of the effective responsivity
sact is valid only for measuring sources that have the same spectral distributions as
the calibration source.
Note 4: Calibrating s E requires the UV radiation to overfill the detector of the meter
head. If the detector is calibrated with respect to radiant power, the effective
receiving area of the meter head and the uniformity of the detector has to be
determined. For many existing meter heads, the cosine correction is very poor. They
may be calibrated and use only in collimated beams of radiation.
4.2. Spectral matching
4.2.1. Spectral mismatch correction factor
According to equation (4.1) a systematic deviation from the estimated value will result,
when a UV meter, which was previously calibrated against source C (calibration source)
with the relatice spectral distribution SC(), is used to measure radiation produced by
another source Z of different spectral distribution SZ(). (see also Note 3).
However, such a deviation can be corrected using a spectral mismatch correction factor,
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
FC,Z, defined as the ratio of the effective responsivity of the meter head with respect to the
radiant quantity of the calibration source sact,C() to the effective responsivity with respect
to the test source sact,Z():

FC , Z
s
 act ,C 
sact , Z
 SC    srel    d
0


 S    A    d
Z

act
0

 S    A    d  S    s    d
C
act
0
Z
(4.3)
rel
0
where SC   , SZ   are the relative spectral distributions of the calibration source C and
test source Z respectively.
Using this spectral mismatch correction factor, the measurement result can be corrected as:
YC  YZ  FC ,Z ,
(4.4)
where
YC
corrected value with respect to calibration source C
YZ
(un-corrected) reading of the radiometer when measuring the test source Z
If no correction is made, the relative deviation f1 of the result, will be
f1 
1
s
1  Z 1
FC , Z
sC
(4.5)
Note 5: f1 is an index which indicates the magnitude of deviation of the results
measured by the meter on different sources. For perfect spectral matching (the
relative spectral responsivity srel() of a meter is identical to the prescribed action
spectrum Aact()), f1=0, regardless of the type of source it is calibrated against or it
measures.
Note 6: f1 is not a unique figure for a meter but associated with the sources specified.
The spectral distributions of some commonly used UV sources are listed in the Appendix
A
4.2.2. Spectral match characteristic f1
For optimising the spectral matching of a radiometer head to a desired action spectrum the
following characteristic should be used,
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS

f1 
s
0
*
rel
( )  Aact    d

 A    d
,
(4.6)
act
0
where

s
*
rel
  
 S   A    d
C
act
0

 S   s    d
C
 srel  
(4.7)
rel
0
Note 7: The definition of f1 of this document is independent of any source spectrum
and therefore can be used for direct comparison of UV meters for a specified
application in terms of spectral matching.
Note 8: For a perfect match (i.e., s rel()=A act() at every wavelength), f1 = 0.
Note 9: The index f1 describes how well the responsivity of the meter matches the
target action spectrum; its value has no relationship with the deviation caused by
spectral mismatch f1.
4.3. Short and long wavelength range response characteristic f u and f r
The out-of-band signal of a UV meter which is the unwanted response of the meter
according to that part of radiation, which lies beyond the action spectrum of interest, is
determined using the following edge filter method. Due to the lack of appropriate short
pass filters, a method using only long pass filters is recommended [2]
While irradiating the UV meter using a polychromatic source with sufficient radiation in
the out-of-band spectral region, the output of the UV meter Y is measured without an
additional filter and using two different steep long-pass filters (output: YL1 and YL2 ). The
first long-pass filters (L1) is chosen that its filter edge coincides with the short-wavelength
edge 1  of the domain of the action spectrum where its transmission should
be  L1 1   0.1 ; the second one, (L2), is chosen such that its filter edge coincides with the
long-wavelength edge  2  of the domain of the action spectrum and its transmission at
that point should be  L2 2   0.1 .
The characteristic of the respective out-of-band response of the meter ( f u , f r ) is
determined as the ratio between the unwanted part of the signal, normalised to the effective
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
transmission of the filter, and magnitude of the in-band signal of the UV meter:
Y
Y  ~L1

fu 
L1
L1
fr 
(4.8)
YL1 YL2

~ ~
L2
YL2
~
L2
(4.9)
YL1 YL2

~ ~
L1
L2
where ~L1 and ~L2 are given by:

~L1 
 S        A    d
L1
0
act
(4.10)

 S    A    d
act
0

 S        d
L2
~L2 
L

 S    d
,
(4.11)
L
where
Y , YL1 , YL2 measured output of the UV meter with and without different filters,
 L1  ,  L21  spectral transmittance of the long-pass filter for determining the out-ofband response,
S  
relative spectral distribution of the lamp used for determining out-ofband response,
Aact  
action spectrum for which the UV meter is designed.
Note 10:
The out-of-band characteristics ( f u , f r ) are determined independently,
i.e. the polychromatic source used for each characteristic may be
different ( S   = S UV   for f u and S   = SIR   for f r respectively).
The source is not defined for this test and also cannot be defined because different action
spectrum will require different source to do this. However, the source used for this method
should have sufficient radiometric output in the outerband regions of the specified action
spectrum. A list of sources and filters applicable for radiometer heads matched for a
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
variety of actinic weighting functions is given in Annex A.
4.4. Directional response
4.4.1. General
The effect of incident radiation on the acceptance area of the radiometer head depends on
the angle of incidence. The directional response function (evaluation of the incident
radiation as a function of the angle of incidence) is determined by the shape and the optical
properties of the acceptance area as well as on the geometric and optical construction of
the radiometer head.
4.4.2. Measurement
For the measurement of directional response, a point radiation source shall be placed in
front of the radiometer head. The distance in-between must be greater than the radiometric
limiting distance, which is a function of source size, detector size and the directional
response of the meter. As a rule of thumb, the minimal distance should be at least 20 times
the largest dimension of either the source or the acceptance area of the meter. If the angle
of radiation of the source is small, much larger distances may necessary Special
precautions should be taken to exclude stray radiation and fluorescence from the
acceptance area of the radiometer head. The radiometer head has to be adjusted with its
optical axis towards the source.
Rotation of the radiometer head around a horizontal or vertical axis varies the angle of
incidence referred to the centre of the acceptance area of the radiometer head. The centre
of revolution should be coincident with the centre of the acceptance area. Measurements of
the signal as a function of the angle of incidence shall be carried out in at least two
mutually perpendicular planes.
Note 11: On adapters using radiation scattering materials, the reference
point for the rotation has to be stated.
Note 12: For radiometer heads with a non-linear relationship between input
quantity and output signal, the measurement should be conducted at
a constant signal level, e.g by changing the irradiance in a defined
way. Otherwise the result must be corrected via the measured inputoutput characteristic of the radiometer head.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
4.4.3. Characterisation for irradiance E
For a UV irradiance meter the systematic deviation in directional response to the incident
radiation is given according to Equ. (4.12),
f 2  ,   
Y  ,  
1 ,
Y   0,    cos 
(4.12)
where
Y(,)
output signal as a function of the angle of incidence , according to
Fig.1.
   


Figure 1. Co-ordinates for evaluation of directional response.
For characterising the directional response, the characteristic f2 according to equation
(4.13) is used,
f2 
85
180
f
*
2
( ,  0)  sin 2 d .
(4.13)
0
NOTE 13: The angle range beyond the limits of integration of equation (4.13),
is excluded, since it leads to results not relevant for practical use.
NOTE 14: According to that definition the characteristic f2 is defined as mean
value of cosine errors averaged over nearly the total half sphere of
reception.
4.4.4. Characterisation for spherical irradiance E0
For a UV meter head in a spherical irradiance meter, the deviation in directional response
is characterised by:
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
f 2,0  ,   
Y  ,  
1
Y   0,   0
(4.14)
For characterising the directional response the characteristic f 2 , 0 according to equation
(4.13) is used:

f 2, 0
1
  f 2,0  ,  0  sin    d
20
(4.15)
4.5. Linearity
4.5.1. General
The input-output characteristic of a radiometer describes the relationship between the
output quantity (reading) and the input quantity.
The linearity of a radiometer is the property whereby the output quantity of the detector is
proportional to the input quantity. In addition to simple proportionality the output of a
linear detector has to be zero, if the input signal is zero.
In general, perfect linearity of a radiometer cannot be assumed.
4.5.2. Measurement
For the measurement of the input - output characteristic and determination of the deviation
from linearity, the radiometer head has to be irradiated continuously or in steps with
defined levels of irradiance. With each level of input quantity, the output quantity (reading)
shall be recorded.
For performing such measurements, certain setups for varying the irradiance on the
radiometer head are mentioned in the literature. The measurements for high levels of
irradiance may also be determined by testing the linearity of the radiometer’s components
(radiometer head, amplifier) separately.
(In general, the method for linearity measurement should follow methods and procedures
recommended by CIE Pub. No…. of TC2-29 and TC2-48)
4.5.3. Characterisation
The function of systematic deviation f3(Y) in linear response of radiometers is given in
equation (4.16),
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
f3 (Y ) 
Y X act, max
s

 1  act  1 ,
Ymax
X act
sact , max
(4.16)
where
Y
output signal due to irradiance on the radiometer head with an input
quantity X act
X act, max
input value corresponding to the maximum output signal Ymax (largest
value of the measurement range)
Ymax
output signal due to irradiation on the radiometer head with the input
Xmax.
sact
effective actinic responsivity of the detector
sact ,max
effective actinic responsivity at maximum irradiance at the given range
The number of measurement points for determining f3(Y) must be chosen sufficiently large
in order to ensure that the transfer function is completely covered.
To characterise the deviation in linearity within the different ranges, the characteristic f3
according to equation (4.17) is used. It equals the maximum value over all measurement
ranges of the function of systematic deviation f3(Y) in equation (4.16).
f 3  f 3 Y max
(4.17)
4.6. Display unit
Since digital display is now universally used in almost all of modern instruments, only this
type of display is considered. For UV meter using analogue displays, the measurement
uncertainty is given by the class of the instrument (classification by IEC 60051)
The accuracy of digital display radiometers is influenced by the relative uncertainty of the
reading and the uncertainty of the quantization step (in general  1 digit). Accordingly, it is
characterised by the quality index f4 defined by equation (4.18),
f 4  f display 
k
k
d ,
Pmax
YB max
YA max
(4.19)
where
fdisplay
TC2-47 Draft Version 3
(4.18)
relative uncertainty of reading, in %
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
k
factor for range changing (e.g. k=10 when the switching of the measuring
range is at the ratio of 1:10)
YB m ax
is the full scale reading of the less sensitive range B;
YA max
is the full scale reading of the more sensitive range A.
Pmax
the number of different output codes possible (e.g. for a 3 1/2 digit
display, Pmax = 1,999)
d
quantization uncertainty.
The characterization and the resulting index f4 from equation (4.18) are determined by the
class of the output apparatus in order to include the largest deviation at the boundary of the
range change.
4.7. Fatigue
4.7.1. General
Fatigue is the reversible temporal change in the responsivity, under constant operating
conditions, caused by incident irradiance.
Note 15: During the operation of radiometers, reversible changes can occur
in the responsivity as well as in the spectral responsivity. These
changes are designated fatigue. Fatigue is generally greater for
higher detector irradiation levels.
4.7.2. Measurement
Fatigue should be measured with temporary stable irradiation at a level close to the highest
measurable one.
The source used and the specific radiometric quantity shall be stated. The operating
conditions (ambient temperature, supply voltage, etc.) should be held constant. The output
signal should be measured as a function of the irradiation period. Before performing the
test, the radiometer head should not be exposed to radiation for at least 24 hours.
4.7.3. Characterisation
For showing the fatigue, the function of systematic deviation f5(t) is given according to
equation (4.20),
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
f 5 t  
Y t 
1,
Y t 0 
(4.20)
where
t
elapsed time since the beginning of irradiating the radiometer head with
the constant irradiance
Y(t)
output signal at time t
Y(t0)
output signal at reference time t0, e.g. t0 = 10 s.
For characterisation, the quality index f5 according to equation (4.21) and the level of
irradiance or radiance used for the test have to be stated,
f5 
Y (30 min)
1,
Y (10 s)
(4.21)
where
Y(30 min) output signal 30 minutes after the beginning of the irradiation
Y(10 s)
output signal 10 seconds after the beginning of the irradiation.
4.8. Temperature
4.8.1. General
Temperature dependence characterises the influence of the ambient temperature on the
absolute responsivity and the relative spectral responsivity of the radiometer. If the
radiometer is operated within an ambient temperature different from that used during
calibration, measurement errors can occur.
4.8.2. Measurement
In order to measure the temperature dependence, the entire radiometer must be exposed to
the desired temperature. The instrument must attain thermal equilibrium before starting the
measurement.
Note 16: In general, it can be assumed that the radiometer will attain thermal
equilibrium at the desired temperature in about one hour.
In case there is a fatigue effect, the radiometer head should be irradiated only during the
measurement. The measurement should be performed at least for ambient temperatures of
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
15 °C, 25 °C (reference temperature) and 40 °C. Radiometers which are used in the field
should also be measured at an ambient temperature of 0°C. The measurement should be
performed at an irradiation level on the radiometer head that approaches the largest value
of an arbitrary measurement range.
4.8.3. Characterisation
The temperature dependence is described by the equation (4.22)
f 6 T , T0  
Y T 
1
Y T0 
(4.22)
where
Y(T)
reading at the ambient temperature T
Y(T0)
reading at the reference temperature T0 = 25 °C
If only one single value should be used for the characterisation of the temperature
dependence, the quality index f 6 has to be used.
f6 
Y T2   Y T1  ΔT

  T1, T2   ΔT
Y T0 
T2  T1
(4.23)
where
T1, T2
range limits of ambient temperature
interior measurements: T2 = 40°C; T1 = 15°C;T0 = 25°C; ΔT = 1°C
field measurements: T2 = 40°C; T1 = 5°C(or 0°C);T0 = 25°C; ΔT = 1°C
Y(Tx)
reading at the ambient temperature Tx
 (T1,T2 )
temperature coefficient
Note 17: The temperature coefficient allows a correction of readings for
other ambient temperature levels.
4.9. Modulated radiation
4.9.1. General
When measuring modulated radiation, the reading of a radiometer can deviate from the
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
arithmetic mean value if the frequency of the modulated radiation is below the lower
frequency limit, and if the peak overload capability is exceeded or if the settling time is not
completed.
The range between the lower frequency limit flow and the upper frequency fhigh of sinusoidal
modulated radiation (modulation degree 1, see figure 2) is determined for a radiometer.
Within this range the meter reading shall not differ more than 5 % from the reading for
unmodulated radiation of the same arithmetic mean value.
The lower frequency limit flow of sinusoidal modulated radiation is the minimum frequency
where the meter reading does not differ more than 5 % from the reading for unmodulated
radiation of the same arithmetic mean.
Erel(t)
The upper frequency limit fhigh of sinusoidal modulated radiation is the maximum
frequency where the meter reading does not differ more than 5 % from the reading for
unmodulated radiation of the same arithmetic mean.
t
Figure 2. Sinusoidally modulated radiation of modulation degree 1.
4.9.2. Measurement
The measurement of the upper and lower frequency limits can be performed by means of a
DC operated source and a rotating sector disc. A homogeneously irradiated acceptance
area is not required. In spectral ranges where applicable, the measurement of the upper and
lower frequency limits can be performed also by means of UV-LEDs, the radiant intensity
of which is modulated sinusoidally by means of a suitable power supply.
Note 18: Suitable means must be employed to ensure that the arithmetic mean
output of the radiation source used for the measurement remains
constant when the modulation frequency is varied.
Note 19: Experience shows that the generation of modulated radiation by
means of a rotating sector disk in combination with a DC powered
lamp can only be used for frequencies up to 104 Hz. Higher
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
irradiances can be achieved by this method, however. The
measuring range should be stated.
The signal level for the measurement of modulated radiation must lead to a reading near
the full scale of the used measuring range.
4.9.3. Characterisation
The behaviour of systematic deviation of the output signal for modulated radiation versus
the frequency, can be characterized using equation (4.24),


Y  
f 7    
 1
 Y   0 Hz  
(4.24)
where
Y   0 Hz  output signal for irradiation with unmodulated radiation
Y  
output signal for irradiation modulated with frequency f, with the same
arithmetic mean value as for steady-state irradiation.
To characterise the effect of modulated radiation using only a single number, the quality
index
f 7  f 7   100 Hz 
(4.25)
is used.
4.10. Polarisation
4.10.1. General
The output signal of a radiometer can depend on the polarisation condition of the measured
radiation. In this case, the output signal Y changes when the linearly polarised quasiparallel
incident radiation is rotated around the direction of incidence.
Note 20: Radiometer heads of irradiance meters may show polarisation
dependence within certain angles of incidence.
4.10.2. Measurement
In order to measure the polarisation dependence, unpolarised radiation from a point
source is required, following the measurement arrangement described in section
4.43.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Note 21: The radiation from a filament source is generally polarised. Depolarisation
can be achieved by placing a silica plate, slightly tilted, in front of the
radiation source. The position of the silica plate, in order to achieve
complete depolarisation, is determined with the aid of a polarisationindependent detector (e.g. a windowless silicon planar photodiode
perpendicular to the incident radiation) placed behind a polarisation filter.
To achieve complete depolarisation of the radiation (including the tilted glass
plate), a polariser (e.g. two sheet polarisers placed back to back with their axes in
parallel) is placed in front of the radiation source. The polariser can be rotated
around the direction of incidence in order to change the position of the plane of
polarisation.
Note 22: Using a second polariser, check whether the first polariser is completely
polarising the transmitted radiation. The maximum Ymax and minimum
Ymin output signals of the radiometer are then measured while rotating the
first polariser. The measurements are to be carried out for at least two
angles , which differ by 90°.
4.10.3. Characterisation
To characterise the polarisation dependence, the characteristic f8() is given
according to equation (4.26),
f 8   
Ymax    Ymin  
.
Ymax    Ymin  
(4.26)
To characterise the polarisation dependence, the quality index f8 is stated for an
angle of incidence  = 30°,  = 0°
f8  f8   30;
  0
(4.27)
4.11. Non uniform radiation
4.11.1. General
The construction of some radiometer heads can lead to a significant non-uniform spatial
responsivity over the acceptance area. With a uniform irradiation on the acceptance area, a
non-uniform responsivity is without influence.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
4.11.2. Measurement
For the measurement, a radiation source is arranged as described in section 4.4.2. A
circular aperture with an inner diameter of 1/10, and an outer diameter of double the size of
the acceptance area is placed in front of the radiometer head’s acceptance area. Stray
radiation should be avoided on the radiometer head.
The circular aperture has to be placed at five positions in front of the radiometer head’s
acceptance area as follows:
Position 1:
clear opening of the circular aperture centred in the middle.
Positions 2 to 5:
centre of clear opening of the circular aperture is placed at a point,
which is located 2/3 along the radius from the centre of the acceptance area. The four
positions (2 to 5) are at 90° intervals around the centre of the entrance aperture.
The output quantities Y1 to Y5 are to be measured at those five positions.
4.11.3. Characterisation
For characterising the influence of non-uniform irradiation on the acceptance area, the
index f9 according to equation. (4.28) is used,
5
f9 
Y
i
i2
 Y1
(4.28)
4  Y1
where Yi is the output signal from the initial irradiation value X i at each of the four
points 2 to 5 in the plane of the entrance aperture of the UV meter head;
Y1 is the output signal from the same initial irradiation value X 1 at the centre of the
entrance aperture.
5
f9 
 Y Y
i 1
i
5Y
,
(4.26)
where Y is dertermined according to Eq. (4.26),
1 5
Y    Yi ,
5 i 1
(4.26)
and Yi output signal at one of the five positions on the acceptance area of the radiometer
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
head with the input quantity X.
4.12. Range change
4.12.1. General
When changing the range of a radiometer to an adjacent range, a systematic deviation may
occur.
Note 23: A deviation in case of range change is often due to careless
calibration in the different ranges.
4.12.2. Measurement
For the measurement of the error arising from a range change, the irradiation on the
radiometer head is to be adjusted to produce a reading of 90 % of full scale in the more
sensitive range A.
The irradiation is then increased by a factor k [refer to equation (4.19)]. When increasing
the irradiation, the range is changed to the less sensitive range B.
Note 24: For radiometers with a linear input - output characteristic (linearity
of the radiometer head), the signal can be simulated by a precision
current source while the radiometer head is disconnected.
4.12.3. Characterisation
The deviation due to range change must be determined for each range. In order to
characterise it, the function f11 A / B according to equation (4.29) is used,
f11  A / B  
YB
1 ,
k  YA
(4.29)
where
YB
reading in the less sensitive range B with the input quantity XB being k
times as large as XA
YA
reading range in the more sensitive range A with the input quantity XA at
about 90 % of full scale
k
factor according to equation (4.19)
XA
input quantity creating the reading YA.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
To characterise the deviation due to range change, the characteristic f11 according to
equation (4.30) is used. It equals the maximum value over all different pairs of
measurement ranges Ai and Bi of the function f11(A/B) in equation (4.29).
f11  f11  Ai / Bi max ;
i  1...n  1
(4.30)
where
n
number of different measurement ranges of the instrument
4.13. Total quality index f total
4.13.1. General
For highly accurate measurement of the actinic radiation of the different UV sources
available, great demands in spectral match of integral measuring devices must be fulfilled.
4.13.2. Characterisation
All single quality indices f i have to be stated by the manufacturer. The total quality index
will then be calculated according to:
f tot  u cal  f1´ f u  f r  f 2  f 3  f 4  f 5  f 6  f 7  f 8  f 9  f11
(4.31)
For special measurement tasks an additional weighting of the single quality indices may be
used. The weighting function used has to be stated.
Note 25:
The quality index f7 for example could be weighted with zero, if the
frequency dependence of the meter for a special measurement task is of
no interest. If a special kind of UV meter is only used for specified field
of views, the quality index f2 for the angular dependence could be
restricted to that given field of view.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
5. Specification of broadband UV –radiometers
5.1. Content of data sheets
For instruments of class L, and A the manufacturer shall provide
 class
 quality indices as listed in Table 1 and Table 2
 the tabulated values of the spectral responsivity with sufficient precision
 the tabulated values the angular response
 the uncertainty of the determination of the spectral responsivity and the quality indices
(only for class L)
of the individual measured and characterized instrument.
For class B and C instruments the manufacturer shall state the class and provide theses
values for a typical instrument identical in construction.
Additionally the relative spectral responsivity may be shown together with the target
function by means of a diagram.
5.2. Defining a class for a radiometer
The UV-radiometers are classified into 4 classes with regard to the quality. The class limits have
been selected to allow the selection of a reasonable class for a certain general application. The
classes are named by the letters L, A, B, C, with the following purpose:
 Class L:
UV-radiometer with highest quality, application under laboratory conditions
 Class A:
UV-radiometer with high quality
 Class B:
UV-radiometer with medium quality
 Class C:
UV-radiometer with low quality
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
5.3. Class limits for irradiance meters
A UV-radiometer, which shall be classified into a certain class, shall not exceed with any of its
indices and its total index the class limits and total limit given for that class in Table 1. The
individual indices shall be measured and calculated using the methods specified in Chapter 4
“Properties of UV-radiometers”.
Table 1:
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Index
Ref.
Symbol 1)
Class L
Class A
Class B
Class C
Calibration
uncertainty
3.1
ucal2)
<0.03
<0.05
<0.10
<0.15
Spectral match
4.2.2
f1 '
UV out of band
response
4.3
fu
 0,002
 0,005
< 0.010
< 0,050
4.3
fr
 0,002
 0,005
< 0.010
Cosine response
4.4.3
f2 3 )
 0,030
 0,060
< 0,090
< 0,200
E0 response
4.4.4
f2,0
Linearity
4.5.3
f3
 0,002
 0,005
< 0,010
< 0,050
Display unit
4.6.3
f4
 0,002
 0,005
< 0,010
< 0,030
Fatigue
4.7.3
f5
 0,010
 0,015
< 0,020
< 0,050
Temperature
dependence
4.8.3
f6
 0,003
 0,005
< 0,010
Modulated light
4.9.3
f7 4 )
 0,002
 0,005
< 0,010
Polarisation
4.10.3
f8
 0,003
< 0,005
< 0,010
4.11.3
f9
 0,040
< 0,080
< 0,150
Range change
4.12.3
f11
 0,004
< 0,010
< 0,020
Total quality
index
4.13.2
ftot
 0,050
 0,100
< 0,150
VIS/IR out of
band response
Non uniform
radiation
< 0,050
< 0,030
< 0,020
< 0,020
< 0,200
< 0,050
< 0,250
Calculation of the total
f tot  u cal  f1´ f u  f r  f 2  f 3  f 4  f 5  f 6  f 7  f 8  f 9  f11
quality index
1) The indices in the shaded fields are used for the calculation of the total quality.
2) Ucal is the expanded measurement uncertainty for the calibration of the UVradiometer combined from the transfer uncertainty and the uncertainty of the
standard for a confidence level of about 95 % (k = 2).
3) The total quality index value of 4 % does not apply for measurements with normal
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Index
Ref.
Symbol 1) Class L
Class A
Class B
Class C
lighting incidence (perpendicular illumination by point sources from a large
distance). In this case the value f2 = 0 shall be used for the calculation of the total
quality index. The maximum value for the total quality index is reduced to 3 %.
4) Measured at 100 Hz.
For measurement on pulsed light sources with small duty cycles care must be taken
not to saturate the photometer response.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
6. Calibration of broadband UV radiometers
The calibration of UV meters can be performed in a number of ways. The method chosen
depend on standards used and the uncertainty required. Hence, the chosen method mainly
depends on the class of the UV-radiometer.
Many UV-radiometers are used under specified conditions, where the type of source to be
measured is defined and therefore, the spectrum will vary only slightly. Only in this
simplified calibration case is it applicable to directly compare the readings of the DUT
(Device Under Test) with those of the standard used under the same irradiation from the
previously defined calibration source. In all other cases the actual spectral distribution of
the source and the effective radiant power or effective irradiance responsivity of the meter
head needs to be measured. In the following, it is assumed that appropriate tests have
already been carried out to characterise the meters as described elsewhere in this
document.
6.1. Measurement of spectral responsivity
In general, the absolute spectral responsivity of the meter head can be calibrated against a
reference detector under a uniform irradiance produced by a tuneable monochromatic
source following CIE Publication No. 64 and CIE Publication from TC2-48
Note 1: If the radiometer head is the build-in type and cannot be separated from
the display unit, then only the spectral responsivity of the meter as a whole unit can
be measured.
Note 2: For calculating spectral mismatch correction factors, only relative spectral
responsivity needs to be measured.
6.2. Measurement of spectral irradiance and spectral distribution
The spectral irradiance produced by an UV source at a defined position can be calibrated
against a spectral-irradiance standard lamp using a spectroradiometer and an appropriate
diffuser or integrating sphere input optic. The method and procedure can be found in CIE
Publication No.63. The spectral irradiance normalised to a selected wavelength (relative
spectral irradiance) is defined as the spectral distribution of the source. Differing from the
visual spectral range, care must be taken during the measurement to avoid stray-light and
fluorescence, especially if integrating spheres are used within the setup to compare
different type of lamp spectra.
Note 1: For calibration of UV meters using detector-based standard, only relative
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
spectral irradiance needs to be measured (see section 7.3).
Note 2: The measurement must be done over a spectral range sufficient to cover the
responsivity range of the detector to be calibrated and should include out of band
measurements.
Note 3: The spectral resolution and scanning interval of the spectroradiometer
should be 1nm or smaller in order to resolve any structure in the source spectrum.
Note 4: Tabulated data may be adopted for a source whose typical spectrum is
included in the CIE spectrum database for the purpose of meter calibration.
6.3. Calibration of effective irradiance responsivity
After having all the values from the above measurements, the calibration of an UV meter
can be performed.
For “primary” calibration of a reference UV-radiometer, the absolute spectral responsivity
s   s0  s rel   of a UV meter has to be determined. With these data the effective
standard irradiance responsivity s act for the standard sources S C   can be calculated
according to equation (4.1) and (4.2):

def
sact 
s0

K
 S  
 S  
0

0
C
srel  d
C
Aact  d
(6.1)
where
S C  
spectral distribution of the standard source
Aact   actinic action spetrum under consideration
srel  
relative spectral responsivity function of the UV meter
s0
absolute value of the responsivity measured in (A/W) of the UV meter at a
chosen suitable wavelength
K
actinic efficacy (normally set to unity)
The additional determination of the effective area of the input aperture can be avoided, if
the absolute spectral responsivity of the radiometer is calibrated by substitution method
against a calibrated irradiance responsivity standard (e.g. a trap detector together with a
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
high precision aperture) in a uniform radiation field.
Note: However, the measurement of the relative responsivities of UV-meters
which are used for high irradiance levels often leads to a serious problem, as
the output power of monochromator based setups, which are used for the
determination of spectral responsivities is often not sufficient. In these cases
tunable laser based system for the calibration of UV-detectors should be
applied
If the effective actinic irradiance of a real source has to be measured with a UV radiometer
calibrated with respect to its effective standard responsivity sact , the output signal of the
radiometer has to be multiplied by the spectral mismatch factor F for this real source.
Eact. 
y
sact.
F
Therefore, the spectral distribution of the real source has to be determined, yielding a
spectral mismatch factor according to equation (4.3):
F
s 2
A2
s1
A2
A1
s 2
 SC    srel    d
 S    A    d
Z

act
(6.3)
 S    A    d  S    s    d
C
A1
act
Z
rel
s1
where
S Z  
spectral distribution of the real source
The limits for the integrals may be adapted to the lower  s1 ,  A1 and upper  s2 ,  A2 limits
of the detector responsivity and actions spectrum, respectively. For the calculation of the
spectral mismatch factor only the relative spectral distributions of the spectra have to be
known.
For industrial use and within calibration laboratories, a spectrally calibrated UV-meter
together with the spectral distribution of a real source, can be used to perform
consecutively secondary calibration of UV-radiometers. The measurement procedure used
for these secondary calibration is based on a substitution method, where the (for a specific
action spectrum) calibrated UV-meter is substituted by the DUT (Device Under Test
suitable for the same action spectrum) in the radiation field of suitable stable source under
constant operating conditions. The irradiance of the source has to be uniform over the
whole detector aperture. As a result the effective spectral responsivity sZ, act of the DUT
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
under the given irradiance condition is:
( s Z, act ) DUT 
(sact ) DUT 
y DUT
y
s
 ( s Z, act ) REF  DUT  act
y REF
y REF. F
y DUT
 sact REF  F *
y REF
(6.4)
(6.5)
with
F* 



0

0

SC  srel,DUT  d   S Z  srel,REF  d
0

S Z  srel,DUT  d   SC  srel,REF  d
(6.9)
0
where F * is nearly equal to unity if the spectral responsivity of the calibrated reference
standard UV-meter is matching the spectral responsivity of the device under test.
In some special cases, the determination of the effective responsivity of a UV-radiometer
according to equation (2.1) may be necessary. If the spectral distribution SZ   and the
absolute irradiance E0 of the source is measured and K set to unity, the effective
responsivity can be calculated using:
s Z ,act 
Y

E0   S Z   Aact  d
(6.2)
0
Different from the determination of the effective standard responsivity, the uncertainty of
this calibration is mainly determined by the knowledge of the absolute spectral irradiance
of the source.
6.4. Calibration standards
6.4.1. Source standards
When method 1 in 7.3 is used, a source standard is required. The source standard to be
chosen should be of the same type and have the same relative spectral distribution as the
sources to be measured by the meter for its end application. It must be able to produce a
stable, uniform and known (calibrated) irradiance at the reference plane of the meter.
Note 1: Most of the spectral irradiance standards are tungsten halogen lamps or
discharge lamps (D2 lamp) with continuous or quasi-continuous spectra, while
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
many industrial UV meters are designed for measuring UV radiation from
discharged sources such as Hg-based discharge lamps which have strong UV lines.
The later types of sources are usually not suitable for use as spectral irradiance
standard due to either its stability or geometry or output uniformity. Therefore,
calibration of UV meters for industrial applications using source standards is
usually not practical and therefore is not recommended.
6.4.2. Detector standards
When method 2 in 7.3 is used, a detector standard for spectral responsivity measurement is
required. UV-enhanced silicon photodiodes have good responsivity, stability and excellent
linearity in the spectral range from 250nm to 1000nm and can be used as such standards.
Note 1: When method 2 is adopted for the calibration, a spectroradiometer
calibrated against a spectral irradiance lamp is needed for the measurement of the
source spectrum, though in this case, only relative measurement is required.
6.4.2.1. Transfer/working standards
Since in principle the spectral responsivity of any UV meter can be directly measured
according to section 6.1 against a reference detector, a specially designed UV meter can be
used as a transfer/working standard meter for the calibration of other meters provided that
it meets the following criteria:
(1) It is designed to measure the same actinic effect as the DUTs. It has well defined
aperture so that its receiving area can be determined accurately as required by
irradiance measurement;
(2) It is of higher class (higher quality) than the DUTs, especially its spectral matching
characteristics;
(3) It has good spatial response characteristic
(4) It has good stability and linearity over a sufficient measuring range;
(5) Its calibration uncertainty is adequate for its application (calibration job)
(6) Its accuracy meets its application requirement (all systematic errors have been
corrected to give correct reading)
(7) It has sufficient spectral responsivity range
A radiometer designed/used as transfer/working standard meter should be fully
characterised according to methods described in Section 4. Its spectral responsivity and
effective irradiance responsivity must be calibrated over a spectral range sufficient to cover
not only the whole designated action spectrum, but also its long- and short- wavelength
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
sides as long as its value is not zero.
Industrial UV meters can be directly calibrated by comparison with a transfer/working
standard under the irradiance of a suitable UV source.
Note 1: A transfer standard meter with precision aperture but no diffuser may be
used with the advantages of better spatial uniformity and clear defined receiving
area. However, this requires the UV radiation used in calibration to be collimated
or quasi-collimated, and is incident at normal angle on the detector surface. It may
also limit its linear range of measurement.
Note 2: A master meter with the same spectral and geometrical (receiving area,
directional and spatial response) properties as the DUTs may be used as the
transfer standard. However, It is preferable that the detector head of the master
meter can be separately calibrated from its electronic/display units. This allows it
to be calibrated at relatively low irradiance level with lower uncertainty. Spectral
responsivity data of the master meter can be used for spectral mismatch correction
for the test meters.
Note 3: Only when such a master meter is used, calibration can be done at a
particular wavelength near the peak of its action spectrum using a quasimonochromatic UV source (e.g. a black light source emitting at 365nm). However,
calibration in such a simplified way may lead to additional error due to the
differences of the master meter and the DUTs in spectral, geometrical and other
properties.
6.5. Calibration conditions
Calibration of UV meters should be performed on an optical bench under dark room
conditions with proper environmental control (23C  2C, relative humidity (5040)%).
Proper safety measures (warning signs, UV protection goggles, gloves, masks etc.) against
UV radiation should be in place.
The following criteria should be met to avoid measurement errors and minimise
uncertainties of calibration:
(1) The calibration source must be of the same type and has the same spectrum as that of
the sources for the end use;
(2) It must have good short-term stability (better than 1 part of 102). Otherwise, a
monitoring detector must be used to compensate any drift of the radiant power during
the changing of detectors in the calibration;
(3) Normal Incidence on detector surface should be used to avoid error caused by nonTC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
cosine directional response of the standard and test meters;
(4) An approximately collimated UV beam with divergence angle less than 5 should be
used for the calibration;
(5) The whole receiving surface of the detectors of both standard and test meters should be
overfilled by the UV beam;
(6) The beam intensity across the detector receiving area should be spatially uniform;
(7) The irradiance level at the reference plane should be sufficient for the calibration.
Note 1: When method 1 is used for calibration, the source standard is also the
irradiating source. The reference plane should be at a distance defined by the source
standard's calibration.
Note 2: Though calibration is done at normal incidence angle, additional test on the
directional response of the test meter can be done to check its performance under uncollimated radiation;
Note 3: Calibration under un-collimated radiation similar to that in the UUT's end use
may be performed using a master meter with the same spectral and geometrical
properties (especially directional response) as the UUTs.
Note 5: A collimated/quasi-collimated beam of higher intensity and good uniformity
may be obtained with the help of suitable optics. However, the use of such optics may
distort the spectrum of the calibration source and result in additional error/uncertainty
to the calibration.
6.6. Calibration interval
The calibration interval of an UV meter should be determined by the nature of the UV
meter (type of filter and detector) especially its aging property, the frequency of its use, the
actual irradiance level and total exposure time. It is well known that the transmittance of
interference filters deteriorates under intense UV radiation. The spectral responsivity of
some many detectors (e.g. silicon photodiodes) also changes under UV radiation and the
change is not uniform but wavelength dependent. As a general rule, the calibration interval
should be shorter if the meter is under heavy use.
6.7. Uncertainty evaluation
6.7.1. Method
The uncertainty associated with the calibration of a UV meter results from the method used
and the model of evaluation. Therefore, it is determined by the uncertainty in the
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
measurement process and the uncertainty of the calibration standards. Each calibration
must have an uncertainty budget, evaluated according to "Guide To The Expression Of
Uncertainty In Measurement" (ISO corrected 1st edition, 1995). The expanded uncertainty
at a confidence level approximately 95% with a coverage factor should be accompanied
with the measured results in the calibration report. Within the calibration report, the
spectral mismatch correction factor FC Z for the different type of standard sources (see
Annex A) for which the UV radiometer could be used and its assigned uncertainty should
be given. Manufacturers should use the value of this expanded uncertainty as a part within
the classification using the quality indices of the UV-radiometer. The calibration
conditions such as the type of the calibration source (the used spectral mismatch correction
factor and its uncertainty), the angle of incidence and level of irradiance must also be
specified. If a working standard meter is used, its calibration uncertainty should be stated
in the calibration report. Procedures for the determination of measurement uncertainties in
the field of radiometry and photometry are given in publication CIE xxx2008/2009 (TC243).
6.7.2. Major sources of uncertainty
The major sources for of measurement uncertainties during calibration are:

Calibration uncertainty of the standard lamp for spectral irradiance in method 1;

Measurement uncertainty of the spectral distribution of the calibration source

Calibration uncertainty of standard detector used for measurement of spectral
responsivity in method 2;

Calibration uncertainty of the working standard meter, if used (method 3);

Drift of the standard lamp used due to aging;

Aging of working standard meter due to changes of the filter transmittance, detector's
spectral responsivity and its electronic circuit

Non-linearity of the standard meter

Measurement uncertainty of electrical quantities (e.g. current measurement from
photodiode) in the standard meter if they are measured separately using a current or
volt meter

Calibration uncertainty of amplifiers

Drift of amplifier

Stability of the irradiation from the UV source
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS

Stray irradiation falling on the detector in the calibration

Positioning uncertainties of both test and standard detector heads

Alignment of the detector heads relative to the beam

Spatial non-uniformity of detector's responsivities

Spatial non-uniformity of irradiating beam

Temperature change of the detector head due to heating by the radiation of calibration
source

Uncertainty caused by the low resolution of some industrial BBUV meters

Uncertainty of time interval measurement for dosage (time integrated) calibration

Random uncertainties (type A) during calibration.
The actual uncertainty evaluation should be made by qualified professionals according to
the method, standards and conditions used for the calibration.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
7. Selection and correct usage of UV radiometers
To avoid errors caused by incorrect selection and usage of UV radiometers, users should
check the following key information and specifications provided by the manufacturer:

End use and target action spectrum

Relative spectral responsivity and spectral mismatch

Source type and spectral distribution

Applied spectral correction factors and their uncertainties
In addition, users should also check other properties of the meter provided by the
manufacturer which are critical for the intended applications such as measuring range, nonlinearity, directional and spatial response, its total quality index, and the calibration
uncertainty underspecified conditions.
7.1. End Use and Target Action Spectrum
The end use of the meter and its target action spectrum should be clearly stated in the
specification.
Note 1: If the UV meter is claimed to be a UVA or UVB meter, the CIE definitions of
these UV bands (UVA: 315nm - 400nm, UVB: 280nm - 315nm) should be adopted and
an action spectrum of a square function is assumed;
Note 2: If the meter is claimed to measure total UV radiation in a specified spectral
range, an action spectrum of a square function over this range is assumed;
Note 3: If the UV meter is to be used for other (actinic) applications with known
action spectrum, the numerical data or reference literature of the action spectrum
should be given.
7.2. Relative spectral responsivity and spectral matching
A plot of the relative spectral responsivity of the meter (over the entire spectral range
in which it is non-zero) together with the target action spectrum should be provided by
the manufacturer. The spectral match characteristic f1' and the spectral mismatch with
respect to a specified standard calibration source (given in Annex A) should also be
included in the specifications of the meter.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Note 1: If only the relative spectral responsivity within the wavelength range of the
target action spectrum is given, then the short- and long-range responses of
the meter should be included by the manufacturer in the specifications.
Note 2: If the meter is intended for general use, numerical data of its relative
spectral responsivity over the entire wavelength range in which its
responsivity is non-zero should be provided, allowing the user to calculate
spectral correction factor for different sources;
7.3. Source type and spectral distribution
The numerical data and a graphic diagram of the spectral distribution (spectrum) of the UV
source to be measured by the meter should be provided by the manufactuer. The type,
model number and name of the manufacturer of the UV source should also be stated in the
specifications. The spectra of ten commonly used UV sources are given in Annex 1 with
tabulated numerical data.
Note 1: The source used for calibration of the meter should have the same
spectrum as that for its intended use, i.e. They must be of the same type and
preferably of the same model and manufacturer.
Note 2: The spectrum of a source should be measured by a standard lab using
suitable standard of spectral-irradiance.
Note 3: If the UV source to be measured can be found in Annex 1, the spectrum
data given in Annex 1 can be accepted;
7.4. Spectral mismatch correction factor
If the meter is designed for measuring more than one type of UV sources, the spectral
mismatch correction factors, F, for different sources should be given.
Note 1: The values of the spectral correction factors may be built in the meter
display unit allowing automatic correction when different sources are measured;
Note 2: If the values of the spectral correction factors are given in the form of
actual data, manual correction should be carried out by the user when measuring
different types of sources other than that used for calibration. Clear instructions
should be given in the users manual on how to make the corrections for actual
measurement;
Note 3: Either actual numerical data or appropriate reference of the source
spectra should be given together with the type, model number and name of
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
manufacturer.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
8. References:
[1] CIE Publication 17.4 (1987): International Lighting Vocabulary (ILV)
[2] DIN 5031-11
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Annex A
Reference spectra
Table A.1 — Classification of type of lamps
Spectral type
see
Table
Tungsten halogen lamp
C.2
Deuterium-lamp
C.3
Iron doped metal halide lamp
C.4
Mercury medium pressure lamp
C.5
Mercury low pressure lamp
C.6
Fluorescence lamp with high content of UV-B
C.7
Fluorescence lamp with high content of UV-A
C.8
Xenon long-arc (high pressure)
C.9
Halide lamp
C.10
Sun sprectra defined according to CIE/IEC 904- 3
For calculations, the given classified standard spectra have to be used. For calibration
measurements sources with comparable spectral distributions have to be used.
Determination of the out of band singnal
Table A.2 — Selection of lamps and filters fort he determination of the out of band signal
due to radiation below or above the action spectrum of interest.
Actinic spectrum
UV-A
UV-B
Source
Filter
Xenon high pressure lamp according to C.9
or
Tungsten halogen lamp according to C.2
Schott WG 320 / 2 mm
Xenon high pressure lamp according to C.9
Schott WG 295 / 2 mm
or
Schott WG 320 / 2 mm
Schott GG 400 / 3 mm
Tungsten halogen lamp according to C.2
UV-C
TC2-47 Draft Version 3
Mercury medium pressure according to C.5
or
Xenon high pressure lamp according to C.9
Schott WG 225 / 2 mm
Schott WG 295 / 2 mm
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
Actinic spectrum
Photosynthesis
Inactivating of microorganism
Source
Filter
Tungsten halogen lamp according to C.2
or
Xenon high pressure lamp according to C.9
Schott WG 400 / 3 mm
Mercury medium pressure according to C.5
or
Xenon high pressure lamp according to C.9
Schott WG 225 / 2 mm
Schott WG 715 / 3 mm
Schott WG 320 / 2 mm
Schott WG 225 / 2 mm
Damage of skin an eyes
(ACGIH)
Xenon high pressure lamp according to C.9 or
Erythema
Iron doped metal halide lamp according to C.4 Schott UG 1 mm
or
Schott GG 400 / 3 mm
Xenon- High pressure lamp according to C.9
Sofortpigmentierung
Iron doped metal halide lamp according to C.4
Mercury medium pressure according to C.5
or
Self-pigmentisation ?
Schott GG 400 / 3 mm
Schott WG 305 / 2 mm
Schott GG 400 / 3 mm
Xenon- High pressure lamp according to C.9
Schott WG 305 / 2 mm
Metal halide lamp according to C.10 or
Psoriasis
Bilirubinisomerisierung
Xenon- High pressure lamp according to C.9
Schott WG 320 / 2 mm
Tungsten halogen lamp according to C.2or
Xenon- High pressure lamp according to C.9
Schott GG 400 / 3 mm
Schott OG 550 / 3 mm
Uncholuria-isomerization?
Photo-keratitis
Photo-conjunctivitis
Mercury – medium pressure according to C.5
or
Xenon- High pressure lamp according to C.9
Ilmasil PN 235 / 2 mm
Mercury – medium pressure according to C.5
or
Xenon- High pressure lamp according to C.9
Schott WG 225 / 2 mm
Schott WG 320 / 1 mm
Schott WG 320 / 2 mm
The selections of radiant sources and filters given above are recommendations. In addition,
sources have to be chosen such that radiation below and above the spectral range is
provided. Filters have to be chosen to meet the requirements given in chapter 4.3.
TC2-47 Draft Version 3
METHODS OF CHARACTERIZATION AND CALIBRATION OF BROAD-BAND UV RADIOMETERS
TC2-47 Draft Version 3