Basics of Spectral

Measurement

Issued by: JETI Technische Instrumente GmbH Jena, May 2005

Basics of Spectral Measurement

Issued by: JETI Technische Instrumente GmbH

Jena, May 2005
 Basics of Spectral Measurement

Introduction
Modern spectroscopic measurement equipment becomes smaller and
more cost effective. Therefore, it is used in more and more new
application fields besides the classical one in analytics. Spectroscopic
methods are applied in research and production for color measurement,
chemical analysis and quality control, pharmaceutical testing, medical
check up, plant growth observation, testing of light emitting devices, food
control, pollution measurement, etc. Currently, they are increasingly used
in process control, e.g. in dye works, chemical plants, semiconductor
industry and electro plating processes.
Basics of Spectral Measurement

Table of Contents
1 Miniaturized Spectrometers 6
1.1 General Set-up 6
1.1.1 Gratings 6
1.1.2 Spectrometer Set-up 7
1.2 Spectrometer Parameters 10
1.2.1 Wavelength Range 10
1.2.2 Resolution 10
1.2.3 Stray Light 11
1.2.4 ADC Resolution 13
1.2.5 Integration Time 13
1.2.6 Spectral sensitivity 14
1.2.7 Wavelength and Intensity Calibration 14
2 Line Arrays for Miniaturized Spectrometers 16
2.1 General Description 16
2.2 Types of Line Arrays 17
2.2.1 CCD Arrays 17
2.2.2 Photodiode Arrays (PDA) 18
2.3 Criteria of Selection for a Specific Application 19
2.4 Aspects of Read out Electronics 24
2.5 Examples of Line Detector Array 26
3 Selected Applications of Array Spectrometers 27
3.1 Color Measurement in General 27
3.2 Color Measurement of Opaque Surfaces 32
3.3 Photometric and Color Measurement of Transmittive Samples 35
3.3.1 Optical Parameters 36
3.3.2 Measuring Geometry 38
3.3.3 Color measurement 40
3.4 Measurement of Self-luminous Objects 41
3.5 Spectral Fluorescence Measurement 47
3.6 Multichannel Spectral Measurement 50
4. Table of Figures 52

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 5

Miniaturized Spectrometers Chapter 1

1 Miniaturized Spectrometers
1.1 General Set-up
The classical spectrometer consists of an input slit, a rotating dispersive element
(prism or grating), an output slit and a single detector. This arrangement allows the
separation of multichromatic radiation into its spectral components. The main
advantages are the high sensitivity and the low stray light. Several drawbacks,
such as the non-parallel measurement, the moveable elements and the space
consuming dimensions, were overcome by the development of array
spectrometers. This kind of spectrometer uses a detector array instead of a single
detector and therefore only needs fixed components. Meanwhile, they are widely
used in PC coupled, stand-alone as well as hand held devices and begin to find
their applications even in the online process measurement.
1.1.1 Gratings
Most of the time gratings are used as the dispersive element in array
spectrometers. They have the advantages of higher dispersion and lower
production costs than prisms. The basic grating equation is as follows:

λ
sin Θ m = sin Θ i + m (1)
d

with θi and θm - angles of the incident as well as the diffracted wave directions to
the normal of the grating surface, λ - wavelength, d - grating period and m - order
of diffraction (integer value, m = 0, ±1, ±2, ...). This equation is valid for reflexion
gratings, whose grooves are perpendicular to the plane of incidence. Gratings in
spectrometers are mostly used in reflectance mode.
The equation shows, that for m ≠ 0 the radiation of different wavelengths is
separated angularly. This effect is called dispersion. Furthermore, each
wavelength is diffracted into different discrete angles according to the order m,
causing an ambiguity. Spectrometers usually use the dispersed light of one order.
Influences of other orders and the non-dispersive zeroth order have to be
suppressed. Fig. 1 shows the angular separation for monochromatic and
polychromatic radiation, caused by a grating.

Fig. 1 Angular separation of mono and polychromatic radiation

6 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Chapter 1 Miniaturized Spectrometers

It demonstrates the overlap between the different orders - the diffraction angle for
m = 1 in the red region coincidences with the corresponding angle of half of the
wavelength for m = 2. The free spectral range (range without overlap) becomes
smaller for higher diffraction orders.
As mentioned above, the order of diffraction can have negative or positive values.
The common convention, also used here, calls the order positive, if the angle of
diffraction exceeds the angle of incidence and lies on the opposite side of the
grating normal.
The grating equation determines the angles of diffraction, but has nothing to do
with the intensity distribution into the different orders. The fraction of light intensity
diffracted in one order, related to the entire incident intensity is called diffraction
efficiency of this order. Symmetric profiles as the sinusoidal shape of the gratings
in fig. 1 are not very effective. The efficiency can be tuned by varying the shape
and depths of the grooves. An increased efficiency is obtained by a saw tooth
profile. The angle of the swots has to satisfy the condition of the regular reflexion
law to reflect the incident beam into the angle of the chosen diffraction order. This
is called blazing and the condition can be met exactly for one wavelength only, the
blaze wavelength. It is practical to lay this wavelength in the region, where the
other components of the system have low effectivity, e.g. the light source or the
detector, to achieve a homogenisation of the system’s performance.
The change of diffraction angle corresponding to a small change in wavelength is
called angular dispersion of a grating. The following equation is obtained by
differentiating the grating equation to the wavelength with fixed angle of incidence:

dΘ m m
= (2)
dλ d ⋅ cos Θ m

The linear dispersion dl/dλ is the product of the exit focal length f of the
spectrometer and its angular dispersion:

dl dΘ m f ⋅m
= f⋅ = (3)
dλ dλ d ⋅ cos Θ m

It defines the extend of the spectrum on the detector array.

1.1.2 Spectrometer Set-up

Array spectrometers consist of an input slit, a grating, possibly other optical
imaging elements, a line detector and a stable housing. Several basic optical
arrangements are used for spectrometers. The basic demand for the design is to
use collimated light for the dispersion - this improves the measurement accuracy.
One can distinguish between plane grating set-ups and concave grating
spectrometers. The latter type combines diffraction and imaging in one element. If
plane gratings are applied, it is necessary to use additional optical elements for
beam shaping, preferably spherical or aspherical mirrors for ray collimation and
focusing.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 7

Miniaturized Spectrometers Chapter 1

The following figures show examples for both types of set-up (so-called mounts).

Fig. 2 Schemes of Czerny-Turner and flat field concave grating spectrograph

The advantages besides the lack of moving parts are the quasi-parallel
measurement of the whole spectrum, which reduces the measuring time. Further
advantages are the robustness and the small dimensions. The flat field
arrangement is further advantageous because of fewer optical elements, which
have to be adjusted. This type of grating offers a plane focal line, which can be
detected by the plane detector array without remarkable focusing errors.
Nevertheless, plane grating spectrometers as the Czerny-Turner show improved
optical properties because of better compensation possibilities of imaging errors.
Usually, the most efficient diffraction order of the grating is detected by the array;
the detection of other orders is avoided by design or suppressed by several means
(filters, ray traps).
One has to say, that array spectrometers show some disadvantages, too: The
integral illumination over the full wavelength range increases the stray light,
compared with monochromatic set-ups. Furthermore, the sensitivity is lower than
of a monochromator, which can even be fitted out with a photo multiplier. The
precision and resolution is normally less than that of laboratory instruments with a
turnable dispersive element.
The first array spectrometers included original gratings, manufactured by
holography (interference of a splitted laser beam) or mechanical ruling.
Holographic gratings show less stray light because of their better surface quality
than those of mechanically proceeded ones. Furthermore, it is possible to produce
holographic gratings with reduced imaging errors by a special design of the
exposure arrangement.
Such master gratings are replicated into epoxy layers (with an aluminium reflexion
layer and a protection coating) to reduce the costs of the spectrometer. A further
price reduction is obtained with a replication of the master gratings by injection
moulding into a plastics material, e.g. PMMA. This technique is useable for less
demanding applications because of a deterioration of the optical properties.

8 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Chapter 1 Miniaturized Spectrometers

The light inside the spectrometer commonly propagates in free space (air or
glass). The Zeiss spectrometer is an example for this design.

Fig. 3 Free space spectrometer MMS UV (Carl Zeiss Jena)

Another kind of spectrometer is based on the propagation in a slab wave guide

(the LIGA spectrometer. This solution offers the possibility of an extreme
miniaturization. This spectrometer, offered by STEAG microParts GmbH
(Germany), is shown in the following figure.

Fig. 4 Waveguide VIS spectrometer (STEAG microParts Dortmund)

Modern array spectrometers are often fitted out with a fiber optic input for a more
convenient adaption to the measuring problem. There are used single multimode
fibers with or without an additional slit or fiber bundles designed as cross section
transformer from cylindrical to rectangular (slit) shape. In this case the input slit
width w is determined by the fiber core diameter.
The detectors used in UV/VIS spectrometers are silicon-based CCD or photodiode
arrays. CCD arrays show a higher sensitivity, but diode arrays offer a much better
dynamics. The proper selection of the detector array for a special application can
improve the performance of the whole system or can just make it possible. Further
information about the function and the application oriented selection of detector
arrays can be obtained from chapter 2 of these basics (Line Arrays for
Miniaturized Spectrometers).

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 9

Miniaturized Spectrometers Chapter 1

1.2 Spectrometer Parameters

1.2.1 Wavelength Range
Array spectrometers are available for wavelengths from the UV (200 nm) and VIS
to IR (...3 µm). Wavelength ranges of commercial array spectrometers are
indicated in the following figure:

Fig. 5 Scheme of partial electromagnetic spectrum and of wavelength ranges of commercial

spectrometers

The size of the spectrometer is determined by the wavelength range in connection

with the desired optical resolution (see below), demanding in a certain focal length.
JETI´s series of specbos instruments currently cover the range of 380 ... 760 nm,
specbos 1000 UV covers 240 ... 480 nm and the JETI speclight series cover 350
... 850 nm.

1.2.2 Resolution
Several definitions are used for the resolution of a spectrometer. One has to
distinguish clearly between the optical or spectral and the digital or pixel
resolution.
The optical resolution ∆λ is defined by the wavelength difference of two peaks
close together in one spectrum and of same intensity, which can be separated.
The dip between the peaks has to reach a minimum of at least 19 %, related to the
maximum intensity. This definition is called the Rayleigh criterion.

Fig. 6 Definition of the optical resolution (Rayleigh and FWHM)

10 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Chapter 1 Miniaturized Spectrometers

Another more practical definition is related to the measured width of a narrow

spectral line. Its measured bandwidth ∆λFWHM gives information about the
broadening of the line. This bandwith amounts to about 4/5 of the resolution
according to the Rayleigh criterion.
The optical resolution is determined by the width of the input slit, the focal length of
the optical system and the dispersion of the grating. ∆λFWHM is inversely
proportional to the linear dispersion dl/dλ and directly proportional to the output slit
width w´:

dλ
∆λ = w'⋅ (4)
dl

This equation can easily be used for the estimation of a spectrometer resolution,
especially with 1:1 imaging, where the width of the input slit equals to the width of
the output slit (w = w´). The smaller the slit width, the higher is the resolution.
However, reduced slit width reduces the optical energy, entering the spectrometer,
too. This can cause sensitivity problems for a spectrometric system. Common
resolution values of miniaturized VIS spectrometers ay in the region of 5 .. 12 nm,
values even below 1 nm are possible. The core diameter of the input fiber
influences the optical resolution, if it acts as the input slit.
A separate parameter is the pixel or digital resolution. This is the spectral
bandwidth, which is detected by one pixel of the array and is determined by the
width of the pixel and the dispersion of the spectrum. Common values are
2 ... 10 nm/pixel. JETI spectrometers are available with 2 to 10 nm/pixel. The
digital resolution is related to the spectral resolution via the imaging properties of
the spectrometer. The ratio of digital to spectral resolution should be ≥ 3 to detect
safely a peak in the spectrum.
1.2.3 Stray Light
Stray light is radiation of false wavelengths, which strikes a pixel of the detector
array. It is caused by imperfections of the grating, dust, reflexion of the
spectrometer housing or errors of the other optical elements. This parameter
influences the precision of a spectroscopic measuring system in a decisive way. It
gets measured by a broadband illumination of the input slit through a color filter
with long pass characteristics. A filter commonly used therefore is the Schott glass
GG 495. The stray light S is the ratio of the transmission in the blocked
wavelength region below the filter edge (τ(420 nm)) to the transmission in the non-
blocked region (τ(600 nm)).

τ (420nm)
S[%] = 100% ⋅ (5)
τ (600nm)

It shows the influence of the light of longer wavelengths, passing the filter, on the
unwanted intensity in the blocked region. A measured spectrum is shown in the
following figure.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 11

Miniaturized Spectrometers Chapter 1

Fig. 7 Measurement of stray light

In case of a low stray light level (as in fig. 7) an additional neutral density filter is
used for the measurement in the non-blocked region, which will be removed for
the measurement in the blocked region.

τ (420nm) ⋅ τ ( NDfilter )
SL[%] = 100% ⋅ (6)
τ (600nm)

A standard measurement method is described in the ASTM standard test method

E 387 (available from www.astm.org). Detailed information about modifications of
the measurement procedure and their influence on the result can be obtained from
the Agilent technical note "Measuring the stray light performance of UV-visible
spectrophotometers" (available at: www.chem.agilent.com/scripts/Literature
Search. asp).
Another often used stray light measuring method uses a monochromatic light
source (e.g. a HeNe laser). The intensities at the laser wavelength and at another
wavelength, several 10 nm away from the laser wavelength are measured. The
ratio of the latter to the former is a measure for the stray light of the system. This
measuring method delivers much better values, but is more distant to the main
practical applications, where a broadband illumination is used. Therefore, the
method, described first, is recommended.
The stray light causes a non-linearity of the signal at lower power levels and thus
limits the measuring range of the system. For example: the signal in the blue
region of a VIS spectrometer depends on the incident intensity in the other regions
of the spectrum. This effect cannot be compensated precisely by mathematical
calculations in the software.
Every specification of stray light has to be given in connection with the used
measuring conditions. One has to keep in mind, that the higher the bandwidth of
the light entering the spectrometer, the higher the measured stray light ratio. The
value, measured with a long pass filter with an edge in the red region of the
spectrum is not comparable with a value, measured with a yellow edge filter (see
the following figure).

12 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Chapter 1 Miniaturized Spectrometers

GG495
RG2

Fig. 8 Stray light measurements with GG 495 and RG 2

The stray light of array spectrometers is mainly higher than of monochromators

due to the lack of an output slit.
1.2.4 ADC Resolution
The analog intensity distribution of the optical spectrum on the detector array has
to be converted pixelwise to a digital signal by an analog digital converter (ADC).
Common electronic resolutions are 12 ... 16 bit (4095 ... 65 535 counts full scale).
These numbers are finally reduced by the dark spectrum of the line array and the
driving of the converter. The ADC resolution should be in a suitable ratio to the
stray light of the spectrometer and to the dynamics of the detector array. The
dynamics D of the entire system is determined by the ADC resolution RADC,
divided by the noise signal Φ.

R ADC
D= (7)
Φ

A 12 bit spectrometer with a noise signal of 4 counts offers a dynamics of about

1000.
1.2.5 Integration Time
The light intensity coupled into a spectrometer and illuminating a pixel, results in a
proportional signal from the AD converter. The exposure time of the light to the
pixel is called integration time and is a main parameter to adjust the ADC signal. A
level between 2/3 and full scale is suited for best measuring results, using the full
dynamics of the system. Off scale peaks or regions in the spectrum have to be
avoided. Common values for the integration time of array spectrometers are 20 ...
5000 ms. Several instruments, as the JETI spectroradiometer, specbos 1200,
specbos 1201, speclight 1000 and speclight 2000, are equipped with an automatic
integration time adaption.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 13

Miniaturized Spectrometers Chapter 1

1.2.6 Spectral sensitivity

The sensitivity of a spectroscopic system E(λ) is another important issue in many
applications, especially in fluorescence detection, and is a main criterion for the
choice of a spectrometer. It is measured in counts/ Ws and means the ADC signal
r [counts] divided by the optical energy P (λ) ⋅ tint [Ws], entering the spectrometer
optical input.
r
E (λ ) = (8)
P(λ ) ⋅ tint

The spectral sensitivity has to be specified for a given wavelength mainly because
of the spectral dependencies of the grating efficiency and the detector sensitivity.
The measurement can be done with bandpass filtered white light illumination or a
monochromator, from which the radiant flux P (λ) is known.
Furthermore, sensitivity data have to be related to the ADC resolution of the read
out electronics. The main uncertainity in sensitivity measurements is caused by
the measuring error of P (λ), coupled into the spectrometer.
The spectral sensitivity can be adapted in certain limits to the application by a
proper selection of the spectrometer’s detector array (see chapter 2 "Line Arrays
for Miniaturized Spectrometers").
1.2.7 Wavelength and Intensity Calibration
The AD converter delivers the spectral signal in counts for each pixel. The pixels
are numbered and these numbers have to be transformed into the corresponding
wavelength. This can be done by a multiorder polynom as follows:

λ ( n ) = k 0 + k1 ⋅ n + k 2 ⋅ n 2 + k 3 ⋅ n 3 + K + k i ⋅ n i (9)

where n is the pixel number, k0 [nm] the wavelength of the first pixel, k1 [nm/pixel]
the pixel resolution and k2, k3, .., ki are the higher order coefficients.
This approximation with a polynom gives a wavelength precision, which has to be
taken into consideration in the application. The choice for the order of the polynom
depends on the non-linear behaviour of the spectrometer. The linear
approximation is sufficient for a LIGA Spectrometer, whereas a higher order
polynom is necessary for the Zeiss MMS.
The calibration can be done by relating the peaks of a suited spectral lamp (e.g.
Hg in the VIS range) to the corresponding pixel and subsequent calculation of the
k-parameters by regression (e.g. in Excel). JETI offers a special fit program for this
purpose. This program can be used to measure the spectra, indicate the peaks,
calculate the k-parameters and check the result. A spline fit is used to get a
subpixel precision.
A easy way to check the calibration of a spectrometer for somebody who has no
special spectral lamp is to use a common fluorescent lamp (F11, possibly on top of
the lab), which has a spectrum as shown in the following diagram. The intense
peaks of 546 and 614 nm can be used for a roughly test of the calibration.

14 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Chapter 1 Miniaturized Spectrometers

Fig. 9 Typical spectrum of a F11 (TL 84) lamp

Something else, sometimes mixed with the wavelength calibration, is the

calibration of the intensity axis. A non-calibrated spectrometer as the JETI
specbos 1000 has an intensity axis (y- axis) in counts or percent. Therefore, only
relative measurements of light sources, e.g. the spectral distribution of LED, are
possible. One has to keep in mind, that the obtained spectrum is weighted with the
instrument function (grating efficiency, transmittance, detector sensitivity).

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 15

Chapter 2 Line Arrays for Miniaturized Spectrometers

2 Line Arrays for Miniaturized Spectrometers

The line array of a spectrometer is, besides the grating, the main parameter
determining part. A suited choice of the array is vital for the proper adaption of the
system to a certain application. The kinds of line arrays and their operation
principles as well as issues for a correct selection are described in this chapter.
2.1 General Description
Line arrays are detectors with several individual readable sensitive areas, so-
called pixels (picture elements), arranged in one straight line. They can be seen as
“black boxes” transforming a spatially light distribution, in case of spectrometers
the spectrum focus line, into a signal voltage or current distributed in time. This
sequential output signal, the video signal, is further proceeded, mainly with an
analog digital conversion as the first step.
The main parameters of line detectors are the following:
• Pixel number – the number of pixels arranged in the line
• Pixel dimensions – the pixel width and height [µm]
• Pixel pitch – the distance between the centers of two pixels [µm]
• Sensitivity – the wavelength dependent ratio of electrical signal output to the
optical signal input [e.g. V/lx⋅s]
• Wavelength range – the range of wavelengths where the detector can “see”
radiation [nm]
• Dark signal – the output signal without illumination of the detector [e.g. mV]
• Saturation exposure – the illumination level, at which the output signal stays
constant with increasing illuminance [e.g. mlx⋅s]
• Linearity range – the illuminance range where the electrical signal output is
proportional to the impinging energy [e.g. mlx⋅s]
• Dynamic range – the range in which the detector is capable of accurately
measuring the input signal [10x]
• Pixel non-uniformity – the output signal difference of the pixels under same
illumination conditions [%]
Furthermore, the following issues are of interest for an application:
• the scheme of the operation pulses (every line detector needs several clocks
for the read out operation)
• the thermal behavior of the array
• the image lag – a property of line detectors to “remember” the illumination of
previous read out’s

16 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Line Arrays for Miniaturized Spectrometers Chapter 2

2.2 Types of Line Arrays

Currently, there exist three different types of line arrays, which are used in
spectrographs: CCD, photodiode and CMOS arrays. They differ in their distinct
production technologies and their parameters.
2.2.1 CCD Arrays
Principle of Operation
CCDs are array detectors with metal-oxyde capacitors (photogates). During the
illumination by the spectrum focal line a charge (electron – hole pairs) is produced
under the gate. A potential well is created by applying a voltage to the gate
electrode. The charge is confined in the well associated with each pixel by the
surrounding zones of higher potential (see detail in fig. 10). The read out of the
array is proceeded by a charge transfer by means of varying gate potentials
according to special clock schemes. The charges of the pixels are transfered
simultaneously to the shift register(s), followed by a sequential transfer to the
output section, where the charge is converted into a proportional voltage. The
node doing this is first set to a reference level (clamp level) and afterwards to the
signal level. The difference is used as the final signal. This technique is called
Correlated Double Sampling (CDS) and allows a significant reduction of the
system noise.

Fig. 10 Operation principle of a CCD array detector

Examples (see table at the end of this chapter)

There exists a wide variety of CCD arrays from several producers. The Sony ILX
series is often applied in spectroscopy. Other types are the TCD 1201D of Toshiba
and the S 703x series of Hamamatsu. The latter has not only one pixel line but
also 122. It can be used in the so-called line-binning mode. The signals of each
pixel row in the direction perpendicular to the spectrum are additionally combined
to create a greater pixel height.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 17

Chapter 2 Line Arrays for Miniaturized Spectrometers

2.2.2 Photodiode Arrays (PDA)

Principle of Operation
Photodiode arrays consist of several photodiodes arranged in a line. The light
energy impinging on a diode generates a photocurrent, which is integrated by an
integration circuitry associated with this pixel. During a sampling period the
sampling capacitor connects to the output of the integrator through an analog
switch.

Fig. 11 Operation principle of photodiode arrays

There exist two different types of photodiode arrays which differ by kind of their
output signal.
Current output arrays supply the recharge current of the depletion layer capacity
as measuring signal. Therefore, an additional integrator is necessary. The relation
between the peak value of this current and the integral is poor, so read out
electronics based on the reading of the peak value are of lower quality. Especially
with low saturation charge, it is difficult to measure the output correctly. A value of
1 pC equals about 107 electrons, with an electronic resolution of 16 bit one LSB is
represented by only 1 000 electrons.
Voltage output arrays have the integrator on board and deliver a photon flux
proportional voltage for the measurement. This causes fewer problems with the
read out electronics.
Examples
Examples are the S 39xx and S 838x series of Hamamatsu as well as the MLX
90255 of Melexis and LF2C of IC Haus.

18 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Line Arrays for Miniaturized Spectrometers Chapter 2

2.2.3. CMOS Arrays

Principle of Operation
CMOS arrays also use MOS structures as pixels like the CCD arrays. The basic
difference is, that the charge-to-voltage conversion takes place directly in the pixel
cell. This conversion consists again of two steps – the photon – electron and the
charge (electron) – voltage conversion. This difference in read out technique has
significant implications for the sensor architecture, capabilities and limitations. The
following figure shows the scheme of a CMOS line array.

Fig. 12 Operation principle of CMOS arrays

CMOS arrays show better sensitivity than CCD’s because it is easier to produce
low power high gain amplifiers in CMOS technology.
Examples
The S 546x series of Hamamatsu and the L series of Reticon are examples for
CMOS detectors.
2.3 Criteria of Selection for a Specific Application
As stated above, there are many parameters to consider for a proper selection of a
line array for a specific application. Furthermore, it is necessary to weight the
significance of the parameters because there is often no ideal choice possible.
Wavelength Range and Sensitivity
The detectable wavelength range of silicon-based arrays extends from 200 to
1100 nm, above this limit the photon energy is lower than the band gap energy.
For longer wavelengths, InGaAs semiconductor material is used (up to 1.7 µm).
Such arrays are relatively expensive. Furthermore, Si line arrays with anti Stokes
Phosphor coatings are available. They are suited for the wavelength range up to
approximately 1600 nm (depending on the material).
If we consider only silicon arrays there are also differences between the different
types of arrays caused by the set-up and technical treatment. Mainly the extend of
sensitivity in the range below 380 nm and above 800 nm differs significantly. The
window material often is a limiting factor in the UV range. Therefore and due to the
avoidance of additional Fresnel reflexes in the system detector, arrays in
spectrometers are often used windowless.
The following figure shows the relative sensitivity of several line arrays.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 19

Chapter 2 Line Arrays for Miniaturized Spectrometers

Fig. 13 Relative sensitivities of selected Si line arrays

The periodic modulation in many sensitivity curves is caused by white light

interference on thin layers in the sensitive area and can be avoided by
antireflexion coatings.
Typical absolute sensitivity values of different line arrays are shown in the
following table:

Line Array Typical Sensitivity

Hamamatsu S546x series 5.3 V/(lx⋅s)
Sony ILX 5x1 series 200 V/(lx⋅s)
TAOS TSL 140x series 25 V/(µJ/cm2 = 171 V/(lx⋅s) @ 550nm
Toshiba TCD 1201D 80 V/(lx⋅s)
Melexis MLX 90255 1.3 V @ 10 µW/cm2 @ 0.7 ms
Reticon L series 2.9⋅10-4 C/(J⋅cm2)

Pixel Dimensions (pitch and height)

The pixel dimensions are an essential criterion for the selection of an array. The
pixel width and pitch influence the digital resolution of a spectrometer and stand in
relation to its optical resolution (see chapter 1.2.2.). Furthermore, it is necessary to
adapt the pixel height to the height of the spectrum image. If the pixel height is too
low, a part of the signal is lost and the efficiency of the system is low. On the other
side a pixel height much larger than the spectrum height ends in an unwanted
increased stray light influence. Some 2 – dimensional arrays allow a line binning
for the virtual creation of a bigger pixel height.

20 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Line Arrays for Miniaturized Spectrometers Chapter 2

The pixel line has to be precisely adjusted to the spectrometer focal line to avoid a
strong thermal drift of the intensity output. Therefore, JETI offers a special light
source JLQ 1, which simultaneously emits three wavelength ranges in the blue,
green and red into one fiber. The three outputs are controllable, this can be used
with a precise mechanical positioning of the array in 3 directions (parallel and
perpendicular to the pixel line as well as the angle between spectrum and pixel
line) for a precise adjustment.
Dark Output
The dark output is a small electrical output of the line array without incident light. It
is caused by thermal generation of carriers in the light sensitive elements, mainly
due to Si – SiO2 interface states. It has a strong correlation with the operation
temperature. The dark output doubles for every temperature increase of 6 ... 10 K.
Therefore, line arrays are cooled in low light level applications, e.g. in astronomic
measurements.
Furthermore, the dark output depends on the integration time. Therefore, it is
necessary in spectroscopic measurements to proceed a dark measurement at the
same integration time which is used for the spectrum measurement. This can be
done by a mechanical shutter (e.g. JETI spectroradiometer specbos 1201) or with
a switched off light source (e.g. JETI colorimeter specbos 4000).
The dark output level of the spectrometer electronics has a relative niveau. It can
be moved by the choice of gain and offset settings.
Saturation and Linearity Range
Saturation exposure is that level of photon intensity where the photon signal of the
detector is no more dependent on the incident light flux. The value depends on the
doping of the detector material. A rule of thumb is that a good linearity is achieved
with photodiode and CCD arrays modulated between 0 and approximately 75 % of
the saturation charge (exposure). Above this value the non-linearity raises
significantly. The following figure shows the linearity behavior of a current and a
voltage output photodiode array.

Fig. 14 Linearity behaviour of two line arrays (Hamamatsu)

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 21

Chapter 2 Line Arrays for Miniaturized Spectrometers

The saturation exposure is given in the array data sheets, [mlx⋅s] and [nJ/cm2] are
used as units, depending if it is seen from a photometric ar a radiometric point of
view.
The following table shows some values:

Line Array Typical Saturation Exposure

Hamamatsu S 390x series 180 mlx⋅s = 26.3 nJ/cm2 @ 550 nm
Hamamatsu S 392x series 220 mlx⋅s = 32.1 nJ/cm2 @ 550 nm
Hamamatsu S 837x series 145/570 mlx⋅s =21.2/ 83.2 nJ/cm2 @ 550 nm
Sony ILX 5x1 series 40 mlx⋅s = 5.8 nJ/cm2 @ 550 nm
TAOS TSL 140x series 92 … 140 nJ/cm2
TAOS TSL 1301 7 nJ/cm2
Reticon L series 35 nJ/cm2
PVs ELIS-1024 800 .. 6400 Ke- full well
(dep. from selected pixel width)

Noise and Dynamics

There are two main categories of sources, which generate noise in a line array.
1. Fixed pattern noise
It consists of the dark noise and read out noise (e.g. due to internal
switching). The fixed pattern noise is constant when the read out conditions
are fixed and the integration time is not changed. So it is possible to
eliminate it by subtraction from the measured signal.
2. Random noise
It is traceable to erroneous fluctuations of voltage, current or electrical
charge, which are caused in the signal output process. If the fixed pattern
noise is subtracted, the random noise determines the lower limit of light
detection or the lower limit of the system dynamic range.
The dynamics is that range in which the detector is capable of an accurately
measurement of the signal and is defined as the maximum detectable signal
divided by the minimum signal level. The maximum detectable signal is limited by
the detector saturation, the minimum signal by the noise of the system.
The dynamics depends on the integration time due to the thermal generation of
noise.
Every kind of photon detectors only reaches a typical dynamics. Standard CCD
arrays, like the Sony ILX 511, show a dynamics of 250 ... 300, related to full scale
driving. If the spectrum has also regions with lower signal levels, the dynamics
becomes even lower in these regions.
The dynamics of photodiode arrays is much higher. It depends on the operation
regime and is determined by the analog multiplexer on chip and the integrators.
Dynamics values of > 3⋅104 can be reached with the Hamamatsu S39XX
photodiode array series at room temperature, if it is used with a low noise

22 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Line Arrays for Miniaturized Spectrometers Chapter 2

integrator amplifier and a subsequent high performance 16 bit ADC. Photodiode

arrays with integrator on board show a lower dynamics in main cases. An essential
condition for a high dynamics is a high saturation charge of the array.
In case of minimum speed demanding applications it is possible to increase the
dynamics by averaging.
Temperature Drift of Sensitivity and Dynamics
There are two effects, which cause a significant temperature dependence of the
detector sensitivity: The absorption coefficient increases slightly with temperature.
This effect increases the IR efficiency, but conversely decreases the UV efficiency.
The second more significant effect is the exponential increase of the thermally
excited electron-hole pairs with temperature. This leakage current doubles every
6 ... 10 K increase in temperature, connected with a raising noise.
Pixel Non-uniformity
The individual pixels of an array have different sensitivities caused by crystal flaws
in the substrate and slight local variations in the wafer process. This uniformity is
measured during a uniform illuminance of the sensitive area (50 % of saturation
exposure) and calculation of the maximum deviation from the average value.
Often the pixel non-uniformity is not of interest, especially if spectrometric systems
are referenced. Thus the specific pixel differences are included in this reference.
Integration Time
The integration time is the main parameter for the user to adapt the modulation of
a spectrometric system to the signal level of the measurement. In many
applications it is necessary to have a linear relation between the integration time
and the signal, especially in radiometric measurements.
Image Lag
The remain of charges on the PD/ CCD capacity or integrator is called image lag.
It causes a virtual subsequent exposure if a dark measurement was done after a
light measurement. Especially at CCD’s, which came into saturation, several
following read out cycles are necessary before the measurement results can be
used.

Fig. 15 Scheme of the image lag

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 23

Chapter 2 Line Arrays for Miniaturized Spectrometers

2.4 Aspects of Read out Electronics

Types of Video Signals
The classical video signal originates from the TV technologies. The defined
voltages are 1 V for the synchronization level, 0.75 V for black and 0 V for white.
This video signal is called negative. Furthermore, all video signals have a positive
offset voltage shifting the levels to higher values. CCD, like standard types of Sony
(ILX series) and Toshiba are designed according to this classic standard.

Fig. 16 Diagram of a negative and a positive video signal

Negative video signals often offer a so-called clamp level, which allows to set back
the integrator to a defined offset voltage. So it is possible to use the clamp signal
for a technique called correlated double sampling (CDS). One capacitor is loaded
with the black level, the other one with the active video signal. The combination of
both capacitors subtracts the charges. The result is a positive video signal for the
ADC. Another solution uses two conversions of the black level and the active
video signal and a following difference operation. This solution is much slower
than the first one.
The AD conversion is mainly based on a positive ADC related to 0 V. Therefore, a
positive video signal offers the advantage of a much simpler subsequent AD
converter. Furthermore, an additional noise due to further amplifiers is avoided.
Main arrays, which have the integrator on board, offer a positive signal. Examples
are the S837X series of Hamamatsu, the TSL series of TAOS and the arrays of
Melexis and IC Haus.
Components of Line Array Read Out Electronics
A read out electronics for line arrays consists of the following main parts:
• Timing: The electronics needs to generate the necessary clock pulses for the
array. These pulses differ from array type to type.
• Drive circuit: It serves for the operation of the shift registers.
• Signal processing circuit: It serves for the CDS and the amplification.
• AD converter: It transforms the analog signal to a digital level.
• Processor: It controls all processes and can be used for a precalculation of raw
data.

24 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Line Arrays for Miniaturized Spectrometers Chapter 2

• RAM memory: There can be stored system data (e.g. the pixel – wavelength fit
– see chapter 1.2.7. or intensity calibration data - see chapter 3.4.) and
measured data (e.g. dark signal or reference signal).
• Interface driver circuits: They manage the data transfer via different interfaces
(e.g. USB, RS 232, parallel, CAN).
Reduction of Image Lag
An avoidance or significant reduction of the image lag effect can be reached by a
continuous read out of the array, also before the start of the real integration time.
JETI’s electronic solution applies a fast scan modus, which is used to remove the
remaining charge carriers in a short time. A disadvantage of this technique is that
the pixels, which are read out first, have a slightly shorter integration time. Another
technique is to read out the array continuously. In this case the measurement will
be distorted after saturation. The fast scan mode results in a „cleaner“ photo
capacitor.

Fig. 17 Read out scheme of a detector array with fast scans

JETI offers several read out electronics for different line arrays. Some types are
suited for more than one detector type, e.g. the electronics, shown in the following
figure. It is suited for line arrays from Hamamatsu, TAOS, Sony, Reticon, Melexis,
Toshiba, iC Haus and Photon Vision Systems.

Fig. 18 Universal controlling electronics versaspec for different photodiode and CCD detector
arrays example Sony ILX511

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 25

Line Arrays for Miniaturized Spectrometers

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Producer Type Pixel area/ number Technology Remarks Specific features
Hamamatsu S390x series 50/ 25 µm x 2500/ 500 µm, N -MOS Current output High UV sensitivity
128/ 256/ 512 (photodiode)
S392x series 50/ 25 µm x 2500/ 500 µm, N -MOS Voltage output High UV sensitivity
128/ 256/ 512/ 1024 (photodiode)
S703 x series 24 µm x 24 µm , CCD Back illuminated Line binning, high
512 x 122, 1024 x 122 sensitivity (fluoresc.
detection)
S837x series 50/ 25 µm x 2500µm, N-MOS High IR sensitivity
128/ 256/ 512/ 1024 (photodiode)
Sony ILX5x1 series 14 µm x 200 µm, CCD Low dynamic range,
2048 and higher but high sensitivity
Examples of Line Detector Array

TAOS TSL140x series 63.5 µm x 55.5 µm, 128 Photodiode Medium dynamic range
TSL 1301 85 µm x 77 µm, 102 Photodiode High sensitivity
(fluoresc. detection)
Toshiba TCD 1201D 14 µm x 200 µm, 2048 CCD Low dynamic range
Melexis MLX 90255 66 µm x 200 µm, 128 Photodiode similar to TSL1401, Integrated charge
higher pixels amplifier
Ret icon L series 50/ 25 µm x 2500 µm, CMOS
128, 256, 512, 1024 (Photodiode)
IC Haus iC-LFF1401 63.5 µm x 56 µm, 128 Photodiode similar to TSL1401
Chapter 2

Photon Vision ELIS-1024 7,8 ... 64,4 µm x 125 µm Photodiode Electronically

Systems selectable pixel number
(128/256/512/1024)
2.5

26
 Basics of Spectral Measurement

3 Selected Applications of Array Spectrometers

Modern array spectrometers as described in chapter 1 find more and more
applications in research labs, production and education. Recent developments
lead to cheaper spectrometers with improved parameters. One example for the
extended application fields are spectral measuring hand held devices. Such
instruments were not possible in the past because of the space demanding
spectrometer units. Furthermore, there are new light sources available, e.g. UV-
LED’s, and the read out electronics became smaller and cheaper, even with
extended performance.
Selected applications, which are related to different JETI products, are described
in the following. These are:
• the color measurement of opaque surfaces
• the photometric and colorimetric measurement of translucent media
• the radiometric and colorimetric measurement of self-luminous samples
• the measurement of fluorescence spectra
• multichannel spectral measurement
The basic rules of color measurement are described in chapter 3.1.
3.1 Color Measurement in General
Color measurement plays an increasing role in quality control of products.
Especially the checking of color differences between reference materials and later
produced samples is important for the customer acceptance. The easiest way is to
observe the samples by the human eye in special cabins, but there arises the
problem of objectivity and reproducibility. So color measuring instruments are
used, which have to imitate the process of human color perception.
"Let us consider the ´classic island experiment´, in which a random collection of
pebbles of all colors are classified by a lonely castaway. Thinking of color in terms
of the common names red, blue, green, etc., her first step is to separate those,
which have color from those, which have not. In other words, she separates the
CROMATIC pebbles from the ACHROMATIC ones. The achromatic pebbles she
now arranges from black through gray to white: i.e. she arranges them in order of
LIGHTNESS. Turning her attention to the chromatic pebbles, she first separates
them into piles of red, yellow, green, etc.: i.e. according to their HUE. Each of
these piles she then sorts by lightness in the same way as for the achromatic
pebbles. However, she notices that there are still pebbles, which appear different
despite being of the same hue and lightness. After some thought, she realizes that
this kind of difference relates to how much the colors differ from gray - in crude
terms, how much color they contain. This third variable is called CHROMA or
SATURATION. Any color can be uniquely specified by the three properties of
HUE, LIGHTNESS and SATURATION." (From: Multi-channel detector
applications. Andor Technology Ltd. Workshop Belfast October 26-29,1992. p.5).
In technical words: Color is a 3 dimensional quantity. The human eye can
distinguish between more than 1 million chromaticity values. Each impression of
color for a human being arises from the superposition of the illumination, the
reflectivity/ transmittance of the object and the detectivity of the human eye (it
contains three kinds of detectors (uvulas) with red, green and blue sensitivity).

27 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2003

Chapter 2 Line Arrays for Miniaturized Spectrometers

Therefore, a color-measuring instrument has to "know" the spectral characteristics

of the illumination and of the eye. The latter characteristics are different from
person to person, so a standard detectivity was defined by the CIE (Commission
International de Eclarage) in 1931 as an average of over 1000 testing persons and
is called the normal observer (a new standard is currently under development). It
consists of the three standardized uvula sensitivities as shown in fig. 18. They are
called values of standard colorimetric observer or color matching functions
x , y and z .

Fig. 1 Functions for the 2° and 10° standard colorimetric observer (available
in tables with 5 nm step width, see e.g. DIN 5033/2 or CIE 15.2, data
downloadable from:
www.hike.te.chiba-u.ac.jp/ikeda/CIE/data/)

It can be seen from the diagram, that the eye sensitivity above 700 nm can be
neglected. This is the reason that some color measuring instruments use only the
range up to 700 nm.
Color perception is a subjective human observation; this makes the measurement
of color very difficult. Influences besides the differing spectral sensitivity of
individual persons are the age, the psychological feeling and the surrounding
conditions, respectively.
The uvulas are distributed across the cornea with different density, resulting in a
changed perception of colors with regard to the illuminated area. The above
curves of fig. 12 show two data sets for a narrow viewing angle (2°) and a wider
one (10°). The detectivity changes slightly for higher angles, so the characteristics
for two observers were defined.
Color measuring instruments can be splitted up into two categories – filter-based
and spectrometer-based devices. Filter devices use three special filters, whose
transmission characteristics are matched to the three functions of the standard
colorimetric observer as precisely as possible. Spectral measuring devices use a
spectrometer and therefore have much more detectors with their sensitivity
distributed across the VIS spectrum. The standard observer characteristics are
used as calculation values, therefore they are more accurate. JETI offers a device
of this type with its specbos 4000. The object is illuminated by an internal light
source and the signal remitted from the surface under test is measured by a
spectrometer.

28 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

There exist three different kinds to receive a color impression for the human eye:
the color, observed in reflexion, e.g. of a car body (A), the color, observed in
transmission, e.g. the colored windows of a church (B) and the color of a light
source, e.g. of a light emitting diode (C). Spectral measuring instruments for A are
commonly called spectro colorimeters for B spectro photometers devices for C
spectro radiometers (see fig. 19).

Fig. 2 Kinds of color impression (the physical terms are A – reflexion

spectroscopy, B – transmission spectroscopy and C – emission
spectroscopy)

Fig. 20 shows a light source illuminating an object, which is observed by the

human eye. The illumination spectrum is remitted by the object and therefore
weighted with its spectral reflexion behavior. The eye receives this signal and
processes it with its own three sensitivity curves. The result is the color impression
of the object, expressed by the tristimulus values X, Y and Z.

Fig. 3 Object with illuminant and observer

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 29

Chapter 2 Line Arrays for Miniaturized Spectrometers

The physical quantity color is a three dimensional value. The following scheme
shows the principal calculation process:

Fig. 4 Procedure of colorimetric calculation (see DIN 5033/2 and /3)

* k is used for the normalization of the tristimulus values like that Y
equals 100 for the pure matt white body

30 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

The tristimulus values of X, Y and Z do not offer information about lightness, hue
and saturation (see the beginning of this chapter). Therefore, they are transformed
into other color systems.
Since the perceived color only depends on the relative amplitudes of X, Y and Z,
the chromaticity coordinates x and y are defined as in fig. 21. Additionally z =
Z/(X+Y+Z). Because of

(X + Y + Z)
x+ y+z = =1 (10)
(X + Y + Z)

only x and y are mentioned. These two values do not give information about the
intensity; therefore it will be extended by Y. The triple xyY is often used to
characterize a color impression.
The x,y diagram has the shape of a sole, the pure wavelengths (spectrum locus)
and the so-called purpur line form its boundaries.

Fig. 5 xy diagram

The intension of color measuring theory during the last decades was to create
systems, which are better adapted to the feeling of the human eye. One of the
mainly used systems is L*a*b* in which visible color differences (∆E*) in the whole
3D space are reflected with approximately the same value. The distinction of the
asterix values (e.g. Y*) to the values without asterix (e.g. Y) is not indicated in
figure 23 for simplicity. The L*a*b* system is better adapted to the subjective color
feeling of the human eye than the xyY system.
As mentioned above, a main task of color measuring is color comparison. The
geometrical distance of the 3 dimensional color values of reference and sample is
used (∆E*) for easier handling of the measuring results.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 31

Chapter 2 Line Arrays for Miniaturized Spectrometers

* *2 *2 *2
∆ E = ∆ L + ∆ a + ∆b (11)

Examples of color measurement can be seen on the demo program JETI specbos
color, which can be downloaded from www.jeti.com/Downloads/SD.html.
3.2 Color Measurement of Opaque Surfaces
Colorimetry (Spectral Reflexion Measurement
The color impression of an opaque body is a result of the scattering of the
illuminating light on colored pigments in the surface region and the following
diffuse escaping of the scattered light. This process is called remission. The
remissioned light is influenced by the illumination and the illuminated surface (see
fig. 20 and 21). Regularly reflected light contains no color information of the
reflecting surface.
The measuring geometry, especially the kind of illumination, has a significant
influence on the measurement. Two standardized geometries are used for the
color impression measurement. The main criterion for the selection of the suited
geometry for an application is the kind of the sample surface. Smooth surfaces,
e.g. of plastics and varnished materials demand a directed illumination by an angle
of 45°, related to the measuring direction. It is preferably arranged cylindrically
symmetric around the perpendicular measurement axis. This arrangement avoids
the influence of gloss on the measuring result (remember: the regular reflex
contains no color information). Rough surfaces, as those of textiles and brickwork,
demand a diffuse illumination, obtained by a lamp arranged in an integrating
sphere. Two kinds of measurement are possible in this case – gloss included and
excluded. The exclusion is obtained by a special gloss trap. The detection of the
remission is done directly in both cases. Schemes of both measuring geometries
are shown in the following figure:

Fig. 6 Both kinds of geometries for reflective measurement of body colors

Left: Directed 45° illumination and 0° measurement (illumination
preferably symmetrically) (45°/ 0°)
Right: Diffuse illumination by an integrating sphere and 8°
measurement (d/8°)

32 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

Light source and detector can be exchanged (Helmholtz reproducibility). This

statement is exactly valid only for samples without fluorescence properties. The
measuring geometry is described by the specification of the illumination, followed
by the specification of the illumination, followed by the specification of the
observation path (according to ASTM and CIE).
The specbos 4000 of JETI has a fiberoptic 45°/ 0° measuring geometry. Its
measuring head has to be placed in direct contact to the object under test.
The results of a color measurement are strongly influenced by the details of the
used measuring geometry, e.g. by the aperture of illuminant and detector. This is
the reason for different results obtained by different instruments, especially if
different manufacturers design them. The main criterion of an instrument is the
comparability between devices of the same type – the device intercomparability.
As outlined in the general chapter about color measurement the color impression
of an object is dependent on the illumination spectrum and on the colorimetric
observer. To obtain comparable measuring results the observer (see fig. 18) as
well as the illumination (e.g. daylight or the light of an incandescent lamp under
specified conditions, see fig. 24) are standardized. Every measuring result has to
include the related measuring conditions.
It is not possible to obtain a standardized illumination spectrum in an ideal way for
measuring purposes, but if the special illumination spectrum of the measuring
device is known, it can be mathematically converted into such standardized
characteristics.
The calculation of the color values is proceeded according to the scheme of fig.
21. The only addition is that the measured spectrum ϕ(λ) is the product of the
standard illuminating spectrum Sλ, which shall be used as reference for the color
coordinates, and the spectral reflexion coefficient (spectral density) of the sample
R(λ)
ϕ (λ ) = S λ ⋅ R(λ ) (12)

The sample spectral density is determined by the ratio of the spectrum measured
with the sample to the spectrum measured with a white standard of known spectral
reflectivity. Therefore, it is not necessary to illuminate with a standard illumination.

I sample ( λ )
R(λ ) = (13),
I ref (λ )

with Isample(λ) the measured spectrum with the sample and Iref(λ) the measured
spectrum with the reference standard. It is a fundamental condition for this
measurement that the instrument illuminant spectrum is kept constant during the
measuring period (measurement of standard and of sample), if no referencing is
used.
The instrument illumination spectrum can have an arbitrary distribution, but has to
include all parts of the visible range from 380 to 760 nm or at least from 400 to 700
nm. Best results are obtained by a homogeneous distribution across the spectrum
due to an equal dynamics.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 33

Chapter 2 Line Arrays for Miniaturized Spectrometers

Fig. 25 shows two different standardized illuminations, which are often used.
Several more spectra, as daylight of other color temperatures, artificial daylight, Xe
lamp light, etc. are defined (see DIN 5033/7).

Fig. 7 Standardized spectra D65 (daylight with a color temperature of 6500

K) and A (incandescent lamp)

Two samples with different remission (or transmission) spectra, which give the
same color values with one illuminant, mainly show different color values for other
illuminants. This effect is called metamery. The value characterizing this behavior,
is the metamery index. It describes the color difference of the samples measured
at a specified illuminant and related to the reference illuminant where the color
difference is zero.
Every body color measurement instrument is delivered with at least one reference
standard. This standard, a white sample, has to be traceable to measurements of
a standardization institute. It is used for calibration of the instrument before a
measurement series to know the instrument illuminant spectrum Iref(λ). So the
measuring results are fitted to this normal and internal effects of the instrument (as
degradation of the light source) are excluded.

Fig. 8 Colorimeter specbos 4000 with calibration standard

34 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

3.3 Photometric and Color Measurement of Transmittive

Samples
Spectralphotometry (Transmission Measurement of Liquids, Filters and
Transparencies)
Modern spectrophotometers, for instance the specbos 3000 of JETI, are easy to
use instruments for non-destructive testing in many application fields. They
basically consist of a regulated light source with collimation optics, a test cell
holder, a focusing optics, a spectrometer, the read out electronics and a
microprocessor for data management and calculations.

Fig. 9 Spectral photometer specbos 3000

The basic measurement of a spectrophotometer is the determination of the

spectral transmission of the sample according to the following formula:

I sample (λ )
τ (λ ) = ⋅ 100% (14)
I ref (λ )

Isample(λ) and Iref(λ) are the measured intensities with sample and with reference at
the different wavelengths λ. The unit of τ(λ) is % or 1. Several application fields
prefer to use the logarithmical expression of absorbance according to the formula

I sample (λ )
A(λ ) = − log = − log τ (λ ) (15)
I ref (λ )

The unit of absorbance is AU (absorbance units). Other physical terms with the
same meaning of absorbance are extinction or optical density.
The main interest of a user in analysis is the concentration of a sample. So it is
necessary to implement the calibration of the concentration – absorbance
relationship into the instrument.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 35

Chapter 2 Line Arrays for Miniaturized Spectrometers

3.3.1 Optical Parameters

The main parameters of a spectrophotometer are:
• wavelength range
• optical resolution
• wavelength accuracy
• wavelength precision
• photometric linearity
• photometric precision
• photometric accuracy
• stray light
In the following there is given an overview of the meaning and test of these
quantities. The issues of wavelength range, optical resolution and stray light are
treated in the general section (see chapter 1.2. Spectrometer Parameter).

Wavelength Accuracy and Precision

The primary wavelength calibration of a spectrophotometer can be done in two
different ways:
• Transmission measurement of media with strong absorption bands of well-
known wavelengths
Commonly used liquids for UV and VIS are solutions of Holmium Oxide and
Samarium perchlorate. They are available in permanently sealed cells. More
convenient standards are filter glasses with strong absorption bands
(Didymium = Neodymium and Praseodymium or Holmium glass). Examples for
theses glasses are BG 36 and BG 20 (Schott). The latter has absorption bands
at 528.7 and 684.3 nm, which are used for the calibration test of the specbos
3000.
• Emission measurement of a line lamp (e.g. low pressure Hg in VIS)
During this calibration only the detecting part of the photometer (the
spectrometer with input optics) is used. The spectral peaks of the lamp are
applied for the calibration see chapter 1.2.7.).

The wavelength-pixel relation is given as a polynom as described in the general

section.
Wavelength accuracy is the deviation of the average wavelength reading at an
absorption or emission band from the known wavelength of this band, while
precision means the ability of a spectrophotometer to reproduce the measured
wavelengths. The test measurement procedures are described in the Standard
Practice E 275 of ASTM (available from www.astm.org).

36 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

Fig. 10 Transmission spectrum of a Schott filter BG 36, thickness 1 mm

Photometric Linearity
It is important to know the range of linearity for the absorbance measurement, and
therefore, in case of analytics the range for a correct concentration measurement.
Potassium dichromate solutions of different concentrations are commonly used as
photometric linearity standard in the UV range. For the VIS range neutral density
filters of different absorbances (e.g. between 0.2 .... 2.5 AU) are used. They are
more practicable and stable, but less precise and more scratch sensitive. With
these standards the determination of the linearity is possible by plotting the
measured absorbance against the known one. Details can be obtained from the
above-mentioned ASTM practice.

Fig. 11 Diagram with different ND filters, useable for photometric linearity

measurement

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 37

Chapter 2 Line Arrays for Miniaturized Spectrometers

The following diagram shows a typical result of such a measurement.

Fig. 12 Photometric linearity of a spectro photometer

It can be seen that the range of linearity is strongly influenced by the stray light
behavior of the spectrometer. The lower the stray light the broader is the
absorbance range, which can be used for a precise measurement.

Photometric Precision
It represents the ability of the photometer to reproduce the result in successive
measurements. The parameter for the photometric precision is the standard
deviation. The value is determined by at least ten transmission measurements of
ND absorbance filters or perforated screens of defined absorbance, followed by
the standard deviation calculation.

Photometric Accuracy
Sometimes, not in all applications, the accuracy of the photometric measurement
is of interest, e.g. if absorption measurements of different labs have to be
compared. The accuracy gets determined with transmission samples, which are
traceable to national standard laboratories, e.g. NIST standard reference
materials. The accuracy is the difference between the true absorbance/
transmittance values and the average of ten measured values.
3.3.2 Measuring Geometry
The measuring geometry with transmission of a collimated beam through the test
sample, as described in the beginning of this chapter is only useful for samples
with low scattering, e.g. colored liquids or colored glasses (0°/0° geometry). In
case of higher scattering media, such as milk or filters with a rough surface, it is
necessary to use a diffuse illumination by means of a lamp in an integrating
sphere and leaving the detection path with a focusing optics (d/ 0° geometry).
Again it is possible to exchange illumination and detection path. A 0°/ 90°
measuring geometry is used for fluorescence applications to reduce the influence
of the exciting beam (see chapter 3.5.).

38 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

Fig. 13 Set-ups for transmission and fluorescence measurement

Liquids are normally measured in test cells. There are several kinds on the market:
round and rectangular, made from plastics (non-returnable), optical glass,
Spectrosil quartz and UV silica. The standard width is 12.5 mm and the optical
path length can vary between 1 and approx. 100 mm. The standard path length is
10 mm. A very important point is the z height dimension, this is the distance from
the bottom of the cell holder to the center of the incident light beam (standard
values are 8.5 or 15 mm, JETI´s specbos 3000 uses 15 mm). The z height
becomes very important when the aperture of the cell is small. For continuous
(online) measurements flow cells with liquid in and output are used.
If test cells are used, the sample of the liquid under test has to be brought to the
instrument. This procedure is not convenient for several applications, e.g. in
process control. In this case dip probes are used, which are inserted into the test
liquid to proceed in situ measurements. Such probes are mainly fiber coupled.
JETIs specbos 3100 is an example for an instrument with dip probe. It has a
fiberoptic output connector for the illumination and an input guiding the signal to
the spectrometer. The dip probe (see Fig. 31) consists of two fibers, which are
linked to these connectors. It furthermore consists of a collimating optics, a
protective window and a mirror with protective layer, housed in a stainless steel
tube. The illumination light, escaping from one fiber, gets collimated, transfers the
distance from the window to the mirrow and backwards, until it is focused into the
second fiber. The optical path length (measuring length) is the double window –
mirror distance (in most of the cases). There are dip probes with fixed (e.g. 1 mm,
2 mm, 5 mm) as well as with variable path length on the market. There exist
several mounting technologies to protect diffusion of test liquid into the probe,
these technologies have strong influence on the price of the probe (e.g. glued,
sealed with O rings or connected by diffusion welding). Furthermore, the
surrounding conditions can be very different (e.g. temperature, pH factor,
pressure), so it has to be checked carefully whether a probe is suited for a certain
application or not.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 39

Chapter 2 Line Arrays for Miniaturized Spectrometers

The influence of a possible fiber movement lies in the 1.. 3 % range, therefore it is
recommended to fix the fibers before proceeding the measurement.

Fig. 14 Fiberoptic dip probe

3.3.3 Color measurement
The color measurement of transmissive media, e.g. liquids or transparent solid
media (color filters, transparencies), has the same physical background as
described for the reflectance measurement. The illumination can be diffuse or
directed and is usually of the C type. The observation is normally under 0° on the
opposite side of the sample (standard EN 1557). The calculation procedure is the
same as in case of reflexion measurement. Therefore, the JETI spectrophotometer
specbos 3000, normally delivered with a spectral measuring software, can be used
with the color measuring software JETI specbos color for such applications. The
only difference is that an aqua dest. filled test cell or an uncolored sheet of
material is used as the reference standard. The spectrum of this reference is set to
100 % for the whole wavelength range or a traceable spectral transmission across
the wavelength range of interest is available. The simplification with the 100 %
reference is especially useful in applications with color comparison only.
In former times, the color determination of liquids was based on a comparison with
master solutions. Based on this, several color scales depending on the
concentration were created for different application fields.
There are the following examples:
• Iod number – describes the color depth of clear liquids as solvents, softening
agents, resins, oil (DIN 6062)
• Hazen number – characterization of roughly water clear liquids, especially with
pale yellow color (ISO 6271)
• Gardner – used for pale yellow samples, e.g. fat (ISO 4630)
• Lovibond number – originally for beer mash characterization, today used in fat
and oil industry (AOCS Cc 13e)
• Seyboldt number – water clear to pale yellow liquids (for pharmaceutical white
oil, paraffin and mineral oil, ASTM D 156)
• Klett number – photometrical value for cosmetics industry

40 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

These numbers are only valid for the same and similar chromaticities. Such color
measuring numbers were defined empirically. Today these color scales are
implemented in the software of spectrophotometers, so it is easy to use the
instruments in different industrial fields. The following figure shows the Iod, Hazen
and Gardner color numbers in the chromaticity diagram for specific test conditions.

Fig. 15 Iod, Hazen and Garner color numbers in the chromaticity diagram
(standard illuminant C, 2° observer, test cell thickness 10 mm)
From: Standard EN 1557

3.4 Measurement of Self-luminous Objects

Spectroradiometry
For the color measurement of light sources like CRT´s and illuminated areas like
projection screens again filter as well as spectrometer based instruments are
used. The calculation of color values is based on the same rules as described in
the general chapter. In contradiction to body color, there are only light source (in
this case as object) and observer (detector) necessary (see fig. 19 C).

Therefore, equation (12) will be replaced by

ϕ (λ ) = S (λ ) (16)

where S(λ) is the spectral power distribution of the light source under test.
Normally, spectroradiometers have a field of view between 1 and 3° to obtain a
small measuring diameter. The measured radiation is focused onto the input of a
spectrometer by a lens or a lens system. The spectrum is used for the calculation
of quantities as radiance, luminance, chromaticity and correlated color
temperature according to the following scheme:

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 41

Chapter 2 Line Arrays for Miniaturized Spectrometers

42 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

Fig. 16 Scheme of spectroradiometric calculation

It can be seen from the second step that it is necessary to calibrate the
spectrometer intensity axis for spectroradiometric measurements to know the
sensitivity of the system. In case of a body color-measuring device it is simply
done with a white reflexion standard, which is delivered with the instrument. The
manufacturer or supplier of the instrument normally does the calibration of a
spectroradiometer. Halogen lamps with an integrating sphere (luminance
standards) and standard lamps for luminous intensity are mainly used. The
recalibration of the instrument is necessary and normally recommended once a
year.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 43

Chapter 2 Line Arrays for Miniaturized Spectrometers

Fig. 17 Spectroradiometer specbos 1200

The color measuring values are the same as at body colors, in particular the
chromaticity values x y or u´v´ are used. Another kind of description for a position
in the chromaticity diagram is a combination of dominant wavelength and color
purity. The following figure shows the determination of both values.

Fig. 18 Chromaticity diagram with indication of the dominant wavelength/

color purity definition and the blackbody locus with Judd lines
The dominant wavelength is determined by the extension of the connecting line
between the location of the equi-energy spectrum E (x = 0.333, y = 0.333) and the
location of the measured chromaticity M (x,y) to the outer border of the diagram,

44 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

the spectrum locus. The wavelength of this point of intersection S is the dominant
wavelength (DW). The ratio between the distance EM to the distance ES indicates
the color purity (PE). The more narrow the spectrum, the higher is the color
purity. Monochromatic sources have a value of 100 %. The locations on the purpur
line have no assigned wavelengths. If the measured color point M lies inside the
triangle 380 nm – E – 780 nm, than the extension is made from M beyond E to the
intersection with the spectral locus. The resulting wavelength is indicated by a
negative sign and called the complementary wavelength.
A further parameter of light sources is the Correlated Color Temperature (CCT).
The chromaticities of black and gray bodies with different temperatures lie on the
blackbody locus in the chromaticity diagram (see fig. 36). This statement is only in
rare cases exactly valid for other spectra. Nevertheless, the color temperature is
an advantageous and simple indication. For chromaticities near the blackbody
locus it is common to give the temperature of the gray body which chromaticity is
nearest to that of the concerning light source. This is the Correlated Color
Temperature (CCT). The CCT tells nothing about the spectrum of the light source
under test. It is the temperature of the blackbody radiator when the color
appearance is the same as the source being tested. So it is not a measurement of
the physical temperature of the light source.
The connecting lines of the chromaticity locations with the same CCT are called
Judd lines. The bigger the distance between a color point and the blackbody locus
the higher is the uncertainty of the CCT. If it is too far away, the information will
become senseless.
The transformation from the chromaticity coordinates to the CCT can be done by
look up tables or, as in case of the JETI spectroradiometer specbos 1100, by
approximation with empirical formula.
The Color Rendering Index (CRI) is a calculation value to specify the color
rendering properties of light sources, based on a test color sample method. It
indicates the color differences of defined (theoretical) samples, illuminated by the
lamp under test and a reference illuminant. 14 samples are defined by their
spectral radiance factors, 8 of them are selected to cover the hue circle, but
moderate in saturation and approximately the same in lightness and the other six
representing a strong red, yellow, green and blue as well as complexion and
follage colors (vary widely in color). CRI is always related to a certain reference
illuminant, which has to be indicated.
The first step of CRI calculation is to determine the set of tristimulus values X, Y
and Z, resulting from the illumination of the defined samples by the light source
under test and the reference illuminant (see fig. 22). Then the corresponding
chromaticity values u and v are calculated, followed by a correction for the
adaptive color shift due to the different state of chromatic adaptation under the
lamp to be tested and under the reference illuminant. After transformation of all
color data into special uniform space coordinates the color differences ∆Ei for all
samples, illuminated with both illuminants, are calculated.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 45

Chapter 2 Line Arrays for Miniaturized Spectrometers

The special CRI is given by a set of Ri according to the formula

Ri = 100 − 4.6 ⋅ ∆Ei with i= 1 ... 14 (17),

whereas the general CRI is the arithmetical means of the eight special CRI for the
test samples 1 ... 8 (see CIE publication 13.3 – 1995, www.cie.co.at).
Furthermore, from the absolute spectral data of light sources integral intensity data
can be calculated. The selection of the proper lighttechnic value depends on the
kind of light source. E.g. the „intensity“ of a LED detected by the human eye is
given as luminous flux (cd), whereas homogeneous sources as CRT screens are
specified by their luminance (cd/m2). The following figure shows the suited
measurement set-ups for different kinds of light sources.

Object Measurement set-up Measurement

Value

Screen, Lumin ance

displays cd
[ m2 ]

filed of view

Segments of
displays cd
[ m2 ]

Directed
bright
Lumin ious intensity
light source
(reflector lamps, [cd]=[ lm]
lensed lamps, sr
LED)
Total radiation of
diffuse
Lumin ious flux
light sources
(bulbs, [lm]
fluorescence
lamps) Integrating sphere

Illuminated
areas Illuminance
(work place),
digital projectors [lx]=[lm2]
m

Diffusor

Fig. 19 Measurement set-ups for different kinds of light sources

46 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

As stated before, it is necessary for all absolute intensity measurements to

calibrate the intensity axis of the spectrometer. This can be done by means of a
lamp in an integrating sphere, whose spectral power data at the output hole are
known. The field of view of the spectroradiometer will be precisely directed into the
hole of the sphere in case of luminance measurement. For illuminance
measurement the spectroradiometer gets a diffusor in front of the optics to create
a cosine detection behavior. The calibration can be done with a light source of
known luminous flux (e.g. incandescent lamp) or luminance (incandescent lamp in
integrating sphere) and corresponding conversion into irradiance/ illuminance.
Spectroradiometers (luminance as well as illuminance meters) are classified
according to definitions to be found in standards (DIN 5032/6 and /7 as well as CIE
publication 69). Several measuring errors, e.g. the deviation from the v(λ) function,
the deviation in directional response, the effect of polarized light, the linearity
deviation, etc. and the maximum ranges of the errors are fixed in these papers.
The instruments are classified into four categories from highest (class L) to lower
precision (class C).
3.5 Spectral Fluorescence Measurement
Photo luminescence is the property of a substance or body to absorb light of a
certain wavelength and emit at a higher wavelength with lower energy and slightly
later. The difference between the exciting and the emitted wavelength is called
Stokes shift. If the process duration lies in the µs or ns range, it is called
fluorescence, in case of a longer time delay between excitation and emission it is
called phosphorescence.
Common applications of fluorescent materials are:
• Characteristic features against counterfeites bank notes and stamps
• Luminous layers in CRT
• Endangering signs
• Optical brightening agents for paper, textiles and polymers
• Analytical applications
The application of such materials makes it necessary to measure their spectral
behavior, e.g. for quality control purposes.
If the reflexion or transmission spectra of non fluorescent media are measured
with a scanning monochromator systems (sequential excitation by several
wavelengths and integral measurement) and line array spectrometers (parallel
measurement), the results are the same. In case of fluorescent materials the
spectra differ. The reflexion/ transmission spectra of a scanning system do not
exceed the 100 % line, because absorption and (fluorescence) emission are
registered combined. If a line array spectrometer (polychromator) with integral
illumination across the whole wavelength range is used, the absorption and
emission behavior is shown correctly. There are regions where the spectrum
exceeds the 100 % line due to fluorescence emission and simultaneous
illumination at the lower excitation wavelength.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 47

Chapter 2 Line Arrays for Miniaturized Spectrometers

The scheme of a spectral fluorescence measuring system is shown in the

following figure. An excitation monochromator is tuned to the absorption band of
the material to be measured, which almost equals the region with most effective
fluorescence excitation. This monochromator can be replaced by a monochromatic
light source (laser or LED with a short pass filter) in case of applications where the
wavelength flexibility is not necessary and the sample is always the same (e.g. in
check of characteristic features against counterfeiting). This solution offers the
possibility of economic systems for QC processes in the production. The sample is
illuminated in reflexion or transmission. A polychromator measures the
fluorescence signal over the full wavelength range. So it is possible to detect
different fluorescence peaks.

Fig. 20 Fluorescence measuring set-up

The differentiation of the reflected/ transmitted and the fluorescence signal is only
possible by a spectral separation on excitation and emission side. This measuring
set-up is called bispectral. Laboratory set-ups consist of two monochromators,
which both are tuned. Such arrangements have a high sensitivity, but a long
measuring time. The result is a three dimensional spectrum (reflexion/
transmission and fluorescence over the wavelength).
The measuring geometries of reflexion set-ups are equal to those in pure reflexion
spectroscopy (see fig. 23). In case of illumination by a sphere the emitted
fluorescence light acts as a secondary source for the probe. Furthermore, glance
effects have to be taken into account. In 45°/ 0° set-ups these effects do not
matter.
The measurement of liquids or dissolved materials is commonly done in special
test cells, suited for fluorescence measurement. The illumination angle is 0°, the
fluorescence signal is measured under an angle of 90° to the illuminant to reduce
the influence of the excitation energy (see fig. 30).

48 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

In fluorescence measurement it is necessary to pay attention to several main

points:
• The absorption band of the sample has to be known or has to be measured (to
know the wavelength for most effective excitation).
• The fluorescence intensity is relatively low, therefore it is necessary to
suppress the influence of the excitation energy on the measuring signal (filters,
geometry).
• The quantum yield of materials is very different, e.g. Rhodamine B has a very
high yield, Eosin and Chlorophyll a substantial lower one.
• Fluorescence analysis is 100 ... 1000 x more sensitive than absorption
spectroscopy. There can be analyzed mixtures with only one fluorescent part
or media with different fluorescent components.
• The fluorescence intensity can be strongly temperature dependent.
Furthermore, it can be quenched, especially by dissolved oxygen.
• The broadband excitation light can cause photochemical reactions, which can
influence the fluorescence intensity,
• Impurities of the sample, of the solvent or the test cell can influence the
fluorescence signal.
• Stray light (Rayleigh) can cause additional peaks, which have to be
distinguished from the fluorescence signal.

The following figure shows an examplary fluorescence spectrum, obtained by the

reflexion measurement of a fluorescence marker in a bank note. Furthermore, the
excitation spectrum (UV LED) as well as the transmission spectra of two blocking
filters are shown.

Fig. 19 Fluorescence spectrum example (Fluorescence marker of a bank note)

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 49

Chapter 2 Line Arrays for Miniaturized Spectrometers

From this short survey can be seen that line array spectrometers show several
advantages in fluorescence measuring applications especially for economic set-
ups for production control.
3.6 Multichannel Spectral Measurement
The simplest multichannel measurement has been used for many years – the
referencing of a light source with a second spectrometer. During the last 10 years
applications with more channels have become more and more of interest.
First, there is to clarify one misunderstanding: Some publications define the
detection of one spectrum as multichannel detection due to the parallel read out of
the pixel of the line array – which contains much more information than available
from a filter instrument. Multichannel in the sense here means the read out of at
least two separate measuring spectra. Application examples are quality control of
light sources during fabrication process, color measurement and pharmaceutical
high throughput measuring instrumentation.
This read out of the spectra can be done in two ways – serial or parallel. Both
kinds show specific advantages and disadvantages.

Fig. 20 Serial and parallel multichannel spectral measurement

Serial read out Parallel read out

Set-up Optical switch (mainly fiber One spectrometer per
switch) + one spectrometer channel
Measuring time long (Channel number * Short (only one integration +
measuring time per channel) read out time)
Calibration/ referencing only one spectrometer Each spectrometer separately
Reproducibility Reproducibility of switch is No mechanical movement –
included in the measuring higher reproducibility
result
Changes of measuring object Same conditions for all
are possible during sequential channels during read out
channel read out
Cost Considerable costs of optical Costs for several
switch spectrometers
Applications Mainly where high end Fast measuring applications
spectrometers are needed

50 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

The JETI product is based on a simultaneous (parallel) measurement with a

separate spectrometer per channel. The costs of miniaturized spectrometers
decreased during the last years, so this solution becomes more and more of
interest. The following figure shows a 9 channel spectroradiometer MCR-09 with the
software displaying all spectra simultaneously This instrument contains a fiber
shutter for an easy dark correction.

Fig. 21 6 channel spectrometer and screen shot of the PC software

Examples for applications of this technical solution are:

• Spectral measurement of multi LED equipped PCB´s
• Long term observation of lamp burn in tests
• Multi angle color measurement

Typical parameters for spectrometric multichannel systems are:

• Channel number 2 ... 12
• Intergration time 10 ms ..... 1 s
• Read out time 4 ... 20 ms
• Repetition rate up to 10 Hz
The read out of the JETI multichannel instruments is proceeded pixelwise. The
single pixel signals from all channels are fitted into each other. So it is guaranteed
that no jitter appears between the spectra.

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 51

Chapter 2 Line Arrays for Miniaturized Spectrometers

Table of Figures
Fig. 1 Angular separation of mono and polychromatic radiation 6
Fig. 2 Schemes of Czerny-Turner and Seya-Namioka mounts 8
Fig. 3 Free space spectrometer MMS UV (Carl Zeiss Jena) 9
Fig. 4 Waveguide VIS spectrometer (STEAG microParts Dortmund) 9
Fig. 5 Scheme of partial electromagnetic spectrum and of wavelength
ranges of commercial spectrometers 10
Fig. 6 Definition of the optical resolution (Rayleigh and FWHM) 10
Fig. 7 Measurement of stray light 12
Fig. 8 Stray light measurements with GG 495 and RG 2 13
Fig. 9 Typical spectrum of a F11 (TL 84) lamp 15
Fig. 10 Operation principle of a CCD array detector 17
Fig. 11 Operation principle of photodiode arrays 18
Fig. 12 Operation principle of CMOS arrays 19
Fig. 13 Relative sensitivities of selected Si line arrays 20
Fig. 14 Linearity behaviour of two line arrays (Hamamatsu) 21
Fig. 15 Scheme of the image lag 23
Fig. 16 Diagram of a negative and a positive video signal 24
Fig. 17 Read out scheme of a detector array with fast scans 25
Fig. 18 Universal controlling electronics SCB 1000e for different
photodiode and CCD detector arrays 25
Fig. 19 Functions for the 2° and 10° standard colorimetric observer
(available in tables with 5 nm step width, see e.g. DIN 5033/2 or
CIE 15.2, data downloadable from: www.hike.te.chiba-
u.ac.jp/ikeda/CIE/data/) 28
Fig. 20 Kinds of color impression (the physical terms are A – reflexion
spectroscopy, B – transmission spectroscopy and C – emission
spectroscopy) 29
Fig. 21 Object with illuminant and observer 29
Fig. 22 Procedure of colorimetric calculation (see DIN 5033/2 and /3) * k
is used for the normalization of the tristimulus values like that Y
equals 100 for the pure matt white body 30
Fig. 23 xy diagram 31
Fig. 24 Both kinds of geometries for reflective measurement of body
colors Left: Directed 45° illumination and 0° measurement
(illumination preferably symmetrically) (45°/ 0°) Right: Diffuse
illumination by an integrating sphere and 8° measurement (d/8°) 32
Fig. 25 Standardized spectra D65 (daylight with a color temperature of
6500 K) and A (incandescent lamp) 34
Fig. 26 Colorimeter specbos 4000 with calibration standard 34
Fig. 27 Spectral photometer specbos 3000 35
Fig. 28 Transmission spectrum of a Schott filter BG 36, thickness 1 mm 37
Fig. 29 Diagram with different ND filters, useable for photometric linearity
measurement 37
Fig. 30 Photometric linearity of a spectro photometer 38
Fig. 31 Set-ups for transmission and fluorescence measurement 39
Fig. 32 Fiberoptic dip probe 40

52 Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005

Basics of Spectral Measurement

Fig. 33 Iod, Hazen and Garner color numbers in the chromaticity

diagram (standard illuminant C, 2° observer, test cell thickness
10 mm) From: Standard EN 1557 41
Fig. 34 Scheme of spectroradiometric calculation 43
Fig. 35 Spectroradiometer specbos 1100 44
Fig. 36 Chromaticity diagram with indication of the dominant wavelength/
color purity definition and the blackbody locus with Judd lines 44
Fig. 37 Measurement set-ups for different kinds of light sources 46
Fig. 38 Fluorescence measuring set-up 48
Fig. 39 Fluorescence spectrum example (Fluorescence marker of a bank
note) 49
Fig. 40 Serial and parallel multichannel spectral measurement 50
Fig. 41 6 channel spectrometer and screen shot of the PC software 51

Basics of Spectral Measurement  JETI Technische Instrumente GmbH 2005 53