Polymer Optical Fibers (POF) have the power to displace
replace traditional communication systems via copper or even glass
fiber in short distances. One main application area is the automotive
industry. There, POF displaces copper step by step because of its lower
weight. Another reason is the nonexisting susceptibility to any kind of
electromagnetic interference. These two advantages render optical
communication systems first choice for the automotive industry.
Furthermore POF offers easy and economical processing and is more
flexible for plug packing compared with glass fiber. POF can be passed
with smaller radius of curvature and without any disruption because of
its larger diameter in comparison to glass fiber.
Another sector where POF applies for communication is the multimedia
in-house Ethernet system, as shown in fig. 1, [1
]. Here different
application scenarios can be applied, which are mainly parted in three
· “A/V Server
(communication between e.g. television, hi-fi-receiver and DVD-player)
· “Control Server
Network” (messaging between e.g. refrigerator and stove)
· “Data Server
Network” (data exchange between e.g. notebook and printer)
All these services and applications provide a large amount of data
which must be carried for communication. Even communication via polymer
optical fiber is limited by 2 Gbit/s.
Hence new ways of data transmission should be found to master these
high bandwidth applications. One promising attempt is to use more than
one wavelength to carry information via optical fiber. This technique
is called Wavelength Division Multiplexing (WDM), [3
There light consisting of various wavelengths is carried simultaneously
over one single optical fiber. Every single monochromatic part of this
propagating light carries information. Hence there is no limitation in
bandwidth for optical fiber using WDM.
But two new parts must be integrated in the communication system. The
first is the Multiplexer which must be placed before the fiber to
integrate every wavelength to a single waveguide. The second component,
the Demultiplexer, is placed after the fiber to regain every discrete
wavelength. Therefore the polychromatic light must be splitted in its
monochromatic parts to regain the information. This technology has the
power to master the bandwidth requirements which are needed to provide
new multimedia applications in various fields of life.
Fig. 1 Local Multimedia
2 WDM Demultiplexer
Each commercial available WDM Demultiplexer performs after
one of the following principles:
a) Arrayed Waveguide Gratings (AWGs),
is only applicable for infrared range and multi-mode fibers.
b) Fiber Bragg Gratings (FBGs) are only
available for infrared range.
c) Thin-Film Interference Filters are
only available for infrared range as well.
The configuration of the new WDM Demultiplexer is shown in fig. 2, [6
Light is carried via a standard step index polymer optical fiber
(SI-POF) with a core diameter of 0,98mm and a cladding thickness of
0.01mm. Therefore the standard POF is 1mm in diameter. The core
material consists of PMMA (polymethylmethacrylat) with a typical
refractive index of nPMMA=1.49 in the visible range. The cladding
consists of fluorinated PMMA with a slightly lower refractive index.
The numerical aperture shows values of 0.5 and hence the emitted light
beam has a divergence angle of 30°.To separate the information
carried by the monochromatic parts of the light, the divergent beam has
to be separated and focussed. In this principal configuration a concave
lens is applied to focus the light. The prism with low reciprocal
dispersive power separates the several colors of light. The goal is to
separate the different wavelengths on the “Detection
in the size of a few millimetres. This separation should be adapted to
an opto-electrical detector, which is situated in the point of focus to
get the information without any cross-talk.
Fig. 2 Principal Sketch
of a WDM Demultiplexer
The sketch shows a basic setup with only three colors: red, green and
blue. There is no limitation in reality, but for the first
configuration it is useful to reduce the transferred wavelengths. This
principle configuration was simulated with the help of computer
simulation software (OpTaliX). One of the early results is shown in
fig. 3. The refraction power to focus the light is divided by two
lenses. The use of two lenses gets better results than the use of one
lens due to aberrations.
Fig. 3 2D Plot of early
A single biconvex lens shows many aberrations, e.g. spherical and
chromatic aberrations. Hence it is more useful to split the refractive
power by two lenses. The result is a lower spherical aberration,
because of the lower radii which are needed with two lenses to achieve
the same refractive power. The chromatic aberrations are reduced by
using plano-convex lenses. A welcome side-effect is produced by the
first lens: collimation of the light. A collimated light beam reduces
the aberrations for a prism. This prism shows a different dispersive
power for different wavelengths. The splitting is higher if the
refractive index is of high value and differs strongly in comparison
with the wavelength. In general at lower wavelengths higher refractive
index are realized and vice versa. The more the gradient of the curve
the better is the separation of every single wavelength. Fig. 4 shows
different characteristics of four typical optical materials of
refractive indexes in relation to the wavelength of the visible
spectrum of light.
Fig. 4 Refractive index
in dependence of wavelength
In fig. 3 layer “0” shows the emitted light of the
can be considered as a point source, if the divergence angel and the
diameter of the core are included. The first lens consists of the
layers “1” and “2”, it is a
to reduce aberrations. The low reciprocal dispersive power prism is
situated between the two lenses and consists of the layers
“3” and “4”. The second
layers “5” and “6”, focuses the
out of the
prism escaping light on a detection layer “7”. On
detection layer, there must be enough space between every single point
of focus to detect the various wavelengths with the help of an
The first results shown are simulated at the very beginning of the
analysis. To underline the result of this configuration a spot diagram
for the detection layer is shown in fig. 5. A spot diagram collects the
transverse aberrations in the image plane resulting from tracing a
rectangular grid of rays (emerging from a single object point) through
the system. As this analysis method shows, the different colors cannot
be separated completely. Only two of the three colors can be separated.
The red color with a wavelength of 660nm and the blue color with a
wavelength of 470nm can be separated only with overlap and high
cross-talk. The green color with a wavelength of 530nm shows the same
Fig. 5 Spot Diagram of
The aberrations of the two lenses and the prism are too strong and
there is no consistent point of focus. The reason of this behaviour is
that the focus is shifted along the optical axis and therefore the
diameter of the spot of every wavelength especially for the red and
blue color is too large.
A second configuration tries to reduce eminently the chromatic
aberrations. The basic difference in the configuration is the use of a
mirror instead of a lens to collimate the divergent light beam. The 2D
Plot is shown in fig. 6. The mirror is a parabolic off-axis mirror. A
parabolic mirror collimates the light to a perfect parallel light beam
emitted by a light source which is situated in the focus point of the
Fig. 6 2D Plot of
A mirror has one main advantage compared to a lens; there is no
chromatic aberration, because the light passes no other material with a
different refractive index. Hence the light caroming the prism is free
of chromatic aberrations. Again layer “0” is the
The improvement and the change of configuration is layer
“1”, the off-axis parabolic mirror. This mirror is
by 90° and therefore the rays hit the concave mirror not
the angular point. The perfect collimated light is separated in its
monochromatic parts with the help of the prism, layers
and “3”. The only lens in this configuration,
“4” and “5”, focuses the rays
detection layer “6”.
The base area of the whole configuration is smaller than 6x10cm2. Hence
it can be considered as a compact solution for a WDM POF Demultiplexer.
This solution should produce better results. In comparison with the
first simulation the spot diagram is shown in fig. 7 as well. One
reason for the better result is the reduced chromatic aberration. The
second is the path length of the rays through this configuration. As
the distance between layers “5” and
detection layer, is increased the gap between the single points of
focus is also increased. The result, as fig. 7 shows, is a gap of about
5mm between the red and the green color and the green and the blue
color. This gap is large enough to detect and regain the information
sent via the POF with the help of a photo-detector.
Fig. 7 Spot Diagram of
Detection Layer of improved configuration
The goal of the project is to develop a new economical way
increase the bandwidth of polymer optical fiber. This is necessary
because of the increasing demand of high-speed communication systems
for e.g. automotive or in-house applications. This demand can be
satisfied with the help of wavelength division multiplexing, where it
is possible to use more than one wavelength to carry information via an
optical fiber. To apply this technique, it is essential to design an
economical Demultiplexer. There are several systems available on the
market, all with one main disadvantage; they are all too expensive for
Hence a new development is shown here. The main function of a
Demultiplexer is the separation of the monochromatic parts of light. It
is exploited that the refractive index of the used material is not a
constant over the full spectral range, but rather depends on the
wavelength of light, as it is shown in fig. 4. Therefore a prism with
low reciprocal dispersive power can easily separate the different
wavelength of light in different directions. This is the core idea for
this Demultiplexer - to use a prism instead of wavelength selective
mirrors or grids.
These presented results show, that it is a good way to design a
Demultiplexer by means of a prism. The first shown configuration has
some improvements in comparison to the principal sketch. The refractive
power to focus the divergent light beam emerging the POF is splitted
into two plano-convex lenses to reduce spherical and chromatic
aberrations. The occurrence of spherical and chromatic aberrations is
so much the worse the stronger the radius of curvature of a convex lens.
To reduce these types of aberration, it is necessary to increase the
radius of curvature, but this causes lower refraction power. Hence to
lower chromatic and spherical aberrations the refraction power is
distributed in two plano-convex lenses. A comparison of spherical
aberrations for different lens forms is shown in fig. 8.
Fig. 8 Spherical
different lens forms: a) simple biconvex lens, b)
“best form”, c) distribution of
refraction power in
two lenses, d) aspheric, almost plano-convex lens 
As the results of the early configuration show, the aberrations are so
strong, that the points of focus are shifted along the optical axis and
therefore the spot-size of the different colors differs extremely in
the detection layer. The spot size is about 0,5mm in diameter for the
blue and the red color. The low reciprocal dispersive power of the
prism is too weak to separate the three colors. Hence only two colors
can be regained. The green color in the middle is overlapped by the red
and by the blue spot.
Therefore a new configuration must be designed to separate the three
colors completely. Two basic attempts must be applied:
a) The first is to reduce the chromatic
and the spherical aberrations again.
b) The second is to optimize the form of
the prism to
achieve better local separation of the applied wavelength.
These two goals are accomplished with the new configuration. To reduce
the spherical and chromatic aberration to a value of zero, an off-axis
parabolic mirror is used instead of a lens. A mirror has one main
advantage, because the light is not passing another medium with a
different refractive index, there cannot be any chromatic aberrations.
To avoid spherical aberration, the characteristic of a parabolic mirror
is exploited. Light emerges the aspheric mirror in a perfect collimated
beam, if the light source is placed in the point of focus of the
mirror. Hence chromatic and spherical aberrations are non-existent.
The second idea to separate every single wavelength is to optimize the
shape of the prism by using different values of angels. If the light is
diffracted stronger the gap on the detection layer between the single
colors increases. These steps increase the gap of every single part of
light dramatically. The gap between the colors is about 5mm in length
(see fig. 7). Hence they can be easily detected by opto-electronic
detectors to process the transferred information. For that size of gap
cross-talk is absolutely negligible (<< 30dB).
Another possibility to gain greater gaps is to optimize the material of
the prism. If the refraction power is stronger and the gradient of the
curve shown in fig. 4 is higher, the results are greater gaps as well.
Hence in the second configuration the prism is made of PC
(polycarbonate, nPC=1.59). And the Abbe number is about 30. The Abbe
number shows the power of dispersion of a material. The lower the Abbe
number the higher is the dispersion and the gradient of the curve shown
in fig. 4.
The prism is applied of a plastic material as well as the fiber. In the
second configuration - in contrast to the first - the lens, which
focuses the light on the detection layer, is made of PMMA. Therefore
every component of the configuration is made of a polymeric material.
This is an enormous advantage because these components can be
fabricated in a very simple and economical process: the injection
molding technique. This manufacturing technology has the power to open
this Demultiplexer for mass market and that is the goal.
As mentioned above, there are several demultiplexing systems available
on the market, but they are all too expensive for most of the
applications shown in the introduction. This configuration shown here
in alliance with injection molding technique can create economical
There are many applications e.g. in the automotive sector
the in-house communication which require communication systems with
high data throughput. These demands grow almost daily. Hence new ways
of data transferring methods must be found to satisfy all application
demands. One auspicious way is to combine the easily manageable and
processable POF technology with the economical injection moulding
technique to use wavelength division multiplexing instead of only
single wavelength technique via optical fiber. Single wavelength
transmission over POF can achieve data rates up to 2Gbit/s. This
limitation can be overcome by several wavelengths carrying information
via the fiber. WDM requires Multiplexers and Demultiplexers.
Demultiplexers can be designed with optical grids or mirrors to
separate the different wavelengths again. These methods are very
expensive and therefore not useable for most applications mentioned
This paper shows a Demultiplexer with a prism. The results show, that
it is possible to design such a configuration. Even the early
simulation shows results that satisfied the demand for a Demultiplexer,
but these results have to be further developed before using them in any
practical application. For that reason the second configuration has
many advancements e.g. an aspheric mirror instead of a lens. These
ameliorations show greater size of gap between every single wavelength
in the detection layer. This causes easy detection for opto-electronic
These results alone are not enough to open WDM over POF for mass
market. Only in combination with polymeric materials for the elements
of the configuration and the fabrication in injection moulding
technology, is it possible to achieve unit prices acceptable for the
broad mass market.
In conclusion, WDM over POF is the solution for the increasing demand
of bandwidth for all fields of applications.
An inexpensive Demultiplexer can be made by means of injection moulding
technique and hence it is possible to use this Demultiplexer in many
applications where high bandwidth is required. The next steps to
develop this demultiplexing technology ready to market are to
manufacture a prototype to approve the simulated results.
We have to thank the State of Saxony-Anhalt and especially the State
Secretary of Education for the “OPTOREF” project
State Excellence Program.
 H. Kragl, “A
product family for simplex POF home networks”, http://www.pofac.de/itg/fg_5_4_1/de/fachgruppentreffen/treffen21.php
 R. T. Chen and G. F. Lipscomb, Eds, “WDM and Photonic
Switching Devices for Network Applications”, Proceedings of
SPIE, vol. 3949, 2000
 J. Colachino, “Mux/DeMux Optical Specifications and
Measurements”, Lightchip Inc.
Lightreading, July 2001
 A. H. Gnauck, A. R. Chraplyvy, R. W. Tkach, J. L. Zyskind, J. W.
Sulhoff, A.J. Lucero, et. al., “One terabit/s transmission
experiment”, in Proc. OFC’96, PD 20, San Jose, CA,
 C. R. Batchellor, B. T. Debney, A. M. Thorley, T. J. B. Swanenburg,
G. Heydt, F. Auracher,et. al., “A coherent multichannel
demonstrator”, Electr. & Comm. Engineer. J., 235- 242
 Multiplex-Sender für Polymerfaserübertragung und
Verfahren zu dessen Herstellung, 10 2005 050 747.6 (Tx) 22.10.2005,
 Demultiplex-Empfänger für
und Verfahren zu dessen Herstellung, 10 2005 050 739.5 (Rx), 22.10.2005
 Multiplex-Transceiver für Polymerfaserübertragung
und Verfahren zu dessen Herstellung, 10 2006 009 365.8 (Trx)
 Thomas Thöniß, “Abbildungsfehler und
Abbildungsleistung optischer Systeme”, www.winlens.de/pdf/papers/Abbildungsfehler.pdf