The present invention relates to an optical filtering device
made from an optical fibre, in particular an optical fibre filter which can be used
in a system for the transmission of wavelength division multiplexing (abbreviated
"WDM") optical signals. The present invention also relates to a process for the
manufacture of this filter, to an optical fibre which can be used to form this filter,
to a system for the transmission of WDM signals using this filter and to a method
for filtering optical signals.
In detail, a WDM optical signal is a (digital or analogic)
signal comprising a plurality of N optical signals which are independent of each
other and each of which has a respective transmission wavelength &lgr;1,
&lgr;2 .... &lgr;N different from that of the other signals.
Each transmission wavelength defines a transmission "channel". Moreover, each signal
has, associated with it, a respective wavelength bandwidth &Dgr;&lgr; of predefined
size - called a "channel bandwidth" or "channel (spectral) size" - which is centred
on the corresponding transmission wavelength. The channel size depends, typically,
on the characteristics of the laser sources used and on the type of modulation used
in order to associate the information to be transmitted with the signal. In the
absence of modulation, typical spectral amplitude values of a signal emitted by
a laser source are in the region of 10 MHz, while in the case of modulation outside
the range of 2.5 Gbit/s they are in the region of 5 GHz.
The WDM signal also has a "spacing between channels" defined
as the wavelength (or - in an equivalent manner - frequency) separation between
the central wavelengths of two adjacent channels. In order to transmit to a high
number of channels in one of the so-called "transmission windows" of the optical
fibres and in a useful amplification bandwidth of the optical amplifiers, the spacing
between the channels of a WDM signal is, typically, in the region of one nanometre.
Generally, in a WDM system, the transmission of signals
occurs in the following manner: the various signals are first generated by respective
optical sources, then multiplexed so as to form a WDM signal, next transmitted along
the same optical fibre transmission line and, finally, demodulated so as to be received
by respective receivers.
In recent wavelength multiplexing optical amplification
and transmission systems (able to transmit, along the same fibre, a very high number
of channels - for example 128 - distributed over a particularly wide spectral bandwidth
- for example 70 nm) and, more generally, in optical signal processing apparatus,
for both instrumentation and sensors, devices made entirely of optical fibre, without
any propagation of the light in free space, are being increasingly used. In particular,
these devices are required for the operations of spectral filtering, multiplexing
and demultiplexing of the channels and separation of the transmission spectrum into
With regard to spectral filtering, it is necessary to use
both devices with a high wavelength selectivity for filtering of individual channels
and wider-bandwidth devices for equalization of the channels in the amplification
bandwidth of the optical amplifiers. Equalization of the channels is necessary since
the gain spectrum of erbium-doped optical fibre amplifiers (which constitute the
most widely used optical amplification means) has a significantly unequalized form
in the region between 1530 and 1560 nm. Despite the progress achieved in the development
of glass matrices for silica-based optical fibres containing various co-dopants
able to "flatten" the spectral gain curve, at the moment silica-based fibres which
have a sufficiently uniform gain profile such as not to require external equalization
are not available.
The configuration most used to form wideband optical gain
modules involves the use of an equalizer filter arranged between two active-fibre
amplification stages. The insertion of the filter between the two amplification
stages has the fundamental advantage of allowing a spectral "redistribution" of
the power available for amplification, instead of simple suppression of the power
in the wavelength regions with a higher gain. The spectral profile of the filter
which offers maximum equalization depends on the operating conditions of the amplifier
(and, therefore, on the power of the pump radiation supplied to each stage) as well
as the number and the wavelength distribution of the channels. In recent systems
where there is the possibility of channel addition/extraction, the number and the
distribution of the channels may change depending on the configuration chosen by
the system manager.
For the abovementioned reasons it has become important
to have optical filters which can be efficiently integrated with the active fibre,
with low insertion losses and with a spectral profile which can be easily modified
depending on the specific use of the individual amplifier.
Different types of filters made directly using optical
fibre are known.
A first type of filter is distinguished by the fact that
the fibre has a portion with a sudden variation in diameter, i.e. a tapered portion.
This region induces, in each signal passing through it, an attenuation which depends
on the wavelength of the said signal. In this way, therefore, spectral filtering
is performed. The spectral form of the attenuation of these filters is substantially
sinusoidal on the wavelength.
Another type of filter, which is known as a "Fabry-Perot"
filter, is formed by an optical fibre and two Bragg gratings formed in the fibre
itself and operating as mirrors so as to define an optical resonator.
More recently so-called long period gratings (LPG) have
been developed, said gratings being distinguished by periodic variations in the
index profile of a fibre (typically by means of exposure to UV radiation) and also
allowing wavelength filtering to be performed.
A further class of filters is that defined by an interferometric
structure of the Mach-Zehnder type. Such a structure must be able to perform separation
of an optical signal into two different distributions of electromagnetic field,
propagate these distributions along respective optical paths into which it is possible
to introduce, in a controlled manner, a mutual delay and, subsequently, combine
again the two electromagnetic field distributions so as to obtain an optical interference
signal, the intensity of which is a function of the wavelength.
Figure 1 shows schematically a Mach-Zehnder filter 50 of
known type, able to operate with two distinct field distributions. This interferometric
structure comprises a first and a second optical fibre 51, 52 joined at two different
points by means of a first and a second fusion coupler 53, 54, for example of the
50/50 (or 3 dB) type. The filter 50 is able to receive at its input a signal Sin
from a first end of the first fibre 51 and provide at its output a filtered signal
Sout to a second end of the first fibre 51. In the section between the
couplers 53, 54, the fibres 51, 52 define optical paths of different length. The
difference in optical path length between the two fibres 51, 52 may be due to the
fact that they have different transmissive properties, so that the signals which
are propagated in one fibre have a different speed from those which are propagated
in the other fibre or, as shown in the figure, may be due to the fact that they
have different lengths L and L+&Dgr;L in the section considered.
The couplers 53, 54 allow power coupling between the electromagnetic
fields which are propagated in the two fibres 51, 52. In particular, the function
of the first coupler 53 is that of exciting two different electromagnetic field
distributions in the optical fibres 51, 52 from the signal Sin. These
electromagnetic fields, which are propagated along different optical paths, accumulate
a relative phase difference &Dgr;&phgr; which is not zero and defined by:
where neff is the effective refractive index of the mode which is propagated
in the fibres, &lgr; is the wavelength and &Dgr;L is the difference in length
between the sections of the two fibres 51, 52 comprised between the two couplers
The second coupler 54 is designed to combine again the
two electromagnetic fields, generating an interference between them which may be
constructive or destructive, depending on the phase shift &Dgr;&phgr; accumulated.
In the simplest case where the fibres 51, 52 are identical
and the couplers 53, 54 have an optical power dividing ratio equal to 50/50 (3 dB
couplers), the optical powers at the two outputs of the second coupler 54, indicated
respectively by P1 and P2, are defined by the following equations:
Figure 2 shows the normalised transmission spectrum T (&lgr;)
of the filter 50 at the output of its first fibre 51, in the case where &Dgr;L
is equal to 5 µm and neff is equal to 1.462. The period of this
curve is not constant and is a function of the characteristics of the waveguides
used. Having a different response for the different wavelengths, the interferometer
may be advantageously used as an optical filter.
A Mach-Zehnder filter such as that described above is,
however, difficult to use in practice, owing to its extreme sensitivity to external
disturbances (for example variations in temperature) and variations in form (in
particular variations in curvature of the fibre). These phenomena cause variations
in the effective refractive index neff and, therefore, in the optical
path, which are generally different for the two fibres. The behaviour of this device,
which is ideally described by the equations (1) and (2), therefore cannot be predicted
precisely in a real situation.
In order to overcome this drawback, a solution which combines
the two waveguides in a single compact structure has been proposed. The US patent
5,295,205 in the name of Corning proposes a filter formed by introducing two optical
fibres which are different from each other inside a glass tube, collapsing the tube
onto the fibres after creating a vacuum inside the tube and, finally, heating and
stretching the tube in two regions located at a distance from each other so as to
form two tapered regions which define modal couplers. The fibres also have different
propagation constants in the zone lying between the two couplers, resulting in a
relative delay between the optical signals propagated therein.
The Applicant considers that this solution is difficult
to realise on account of the technological complexity of certain steps in the production
process, in particular the operations for collapsing the glass tube around the fibres
after creating a vacuum in the tube and forming the couplers at a distance from
each other determined on the basis of the desired spectral form and independently
of the geometry of the tapered region.
An alternative method of producing a Mach-Zehnder interferometer
is that described in international patent application WO00/00860 in the name of
Corning. This document describes a coaxial optical device comprising an optical
fibre and a coupling regulator integral with the optical fibre. The optical fibre
is single-mode in the third spectral window of optical telecommunications and a
glass tube with a refractive index lower than that of the cladding is collapsed
onto the fibre, as described in the already mentioned US patent 5,295,205. In the
region where the collapsed tube is present, the refractive index profile is modified
so as to allow locally the transmission of two modes, in particular the modes LP01
and LP02. These modes, which are mutually perpendicular by definition,
define two distinct field distributions which, as they are propagated, accumulate
a relative phase difference &Dgr;&phgr;. In the region occupied by the glass
tube, non-adiabatic tapered zones able to induce power coupling between the modes
are formed. The tapered zones are formed by means of the normal technique for manufacturing
fusion couplers, by causing sudden reductions in the diameter of the fibre and the
tube collapsed onto it, such as to obtain coupling between the symmetrical modes
LP01 and LP02, but avoid coupling with the mode LP03.
The Applicant also notes that the device described above
requires the execution of technologically complex manufacturing steps, such as collapsing
of a glass tube, under vacuum, onto an optical fibre and the formation of non-adiabatic
tapered zones such as to have a high value of the coupling factor between the symmetrical
modes LP01 and LP02, but without exciting other higher symmetrical
modes such as the mode LP03 (where "coupling factor" or "splitting ratio"
is understood in this case as being the ratio between the power transferred to the
mode LP02 and the remaining power in the mode LP01) .
The Applicant therefore notes that the Mach-Zehnder optical
fibre filters of the known type are made using complex technology which does not
allow easy control of the filter parameters. The critical nature of the manufacturing
process therefore results in high costs and fairly low production outputs.
The Applicant has considered the problem of providing a
Mach-Zehnder optical fibre filter which is easy to produce, compact and has a high
The Applicant has found that a Mach-Zehnder interferometer
which is easy and inexpensive to manufacture and has predetermined spectral characteristics
may be made using a dual-mode fibre designed to allow propagation of the fundamental
mode LP01 and the asymmetrical mode LP11 and provided with
two modal coupling regions (for coupling the modes LP01 and LP11)
in which the refractive index profile is asymmetrical due to the presence of a cladding
zone with a higher refractive index. This zone defines essentially, viewed in cross
section, an annular sector of the cladding in a region adjacent to the core and
has a radial extension corresponding substantially to that of the mode LP11.
The Applicant has found that a filter with coupling regions
of this type may be made using an optical fibre having the innermost region of the
cladding doped so as to provide it with high thermo-refractive properties and by
thermally stressing this region so as to produce the desired asymmetrical and localised
variation in the index profile. This doping may be performed with germanium, phosphorus
and fluorine and must be such that the fibre is able to respond to a thermal stress
of suitable intensity with a variation in the refractive index greater than 5.10-4,
preferably greater than or equal to 10-3, more preferably greater than
or equal to 2-10-3.
The Applicant has also found that the thermal stressing
may be performed by means of the electric arc of a fusion jointer. The Applicant
has found that this technique is particular simple and flexible and may be used
to produce very localised disturbances in the cross section of the optical fibre.
According to a first aspect, the present invention relates
to an optical fibre filter comprising:
- an optical fibre which includes a core and a cladding and through which an optical
signal can pass;
- a pair of coupling regions formed in said optical fibre at a predefined mutual
distance, for producing a power transfer between a first and a second propagation
mode of said optical signal;
- a phase shift region, defined by a section of said fibre lying between said
coupling regions, for producing a phase shift between said first and said second
in which, in said coupling regions, said optical fibre has, in cross section, an
asymmetrical refractive index profile.
Preferably, in each of said coupling regions, said cladding
has, in cross section, an annular sector in which the refractive index is greater
than that of the remainder of said cladding.
The cladding has, in cross section, an inner annular region
adjacent to the core and an outer annular region, said annular sector preferably
belonging to said inner annular region.
The angular sectors of said coupling regions preferably
have substantially the same angular position.
The inner annular region has an internal radius r1
and an external radius r2=k·r1, in which k is preferably between
2 and 6.
Outside of said coupling regions, said optical fibre has,
in cross section, a refractive index profile preferably of the step index type.
Said optical signal has a wavelength comprised in a predefined
transmission wavelength band and said optical fibre is preferably dual-mode in said
wavelength band. Moreover, the filter comprises a first and a second optical connection
fibres which are single-mode in said wavelength band and connected to opposite ends
of said optical fibre.
Advantageously, the inner annular region comprises silica
and oxides of the following elements: germanium, phosphorus and fluorine.
The filter preferably comprises a further pair of coupling
regions formed in said optical fibre, in each of which said cladding has, in cross
section, a further annular sector in which the refractive index is greater than
that of the remainder of said cladding, said further annular sectors having substantially
the same angular position, different from that of said angular sectors.
Preferably, each coupling region of said further pair of
coupling regions is formed in the vicinity of a respective coupling region of said
pair of coupling regions.
Advantageously, the difference between the refractive index
in said annular sector and the refractive index of the remainder of said cladding
is equal to at least 5.10-4 and, more preferably, is equal to at least
The filter according to the present invention may comprise
a plurality of filters as defined above, connected in series.
According to a further aspect, the present invention relates
to an optical fibre which can be used for producing a filter as defined above, comprising
a core and a cladding, the cladding having a radially inner region adjacent to the
core and a radially outer region, in which said radially inner region has a composition
such as to obtain a variation in refractive index equal to at least 5.10-4
following thermal stressing and in which said optical fibre is dual-mode in a wavelength
band lying between 1500 nm and 1650 nm. Preferably, said variation in refractive
index is equal to at least 1·10-3 and, more preferably, is equal
to at least 2·10-3.
The difference n2-n3 between the
refractive index n2 in said radially inner region and the refractive
index n3 in said radially outer region is preferably between +1·10-3
Said radially inner region preferably comprises silica
and oxides of the following elements: germanium, phosphorus and fluorine. Advantageously,
in said radially inner region, the germanium has a concentration of between 2% and
5%, the phosphorus has a concentration of between 0.5% and 2% and the fluorine has
a concentration of between 1% and 2%.
Preferably, the core comprises silica and at least one
element selected from germanium and phosphorus.
Preferably, the fibre is dual-mode in a wavelength band
of between 1500 nm and 1650 nm.
Said inner annular region has an internal radius r1
and an external radius r2=k·r1, in which k is preferably
between 2 and 6.
According to another aspect, the present invention relates
to a process for the production of an optical filter from an optical fibre as defined
above, comprising the step of applying an electric arc to a first and a second portions
of said optical fibre in such a way as to stress the cladding of said optical fibre
thermally in an asymmetrical manner.
Advantageously, said electric arc is generated between
a pair of electrodes and the process comprises the step of displacing said optical
fibre in a controlled manner relative to said electrodes after applying the electric
arc to said first portion and before applying the electric arc to said second portion.
Said electric arc has a duration preferably less than 400
ms, and more preferably less than 300 ms, and has a current intensity preferably
of between 8 and 14 mA and more preferably between 10 and 11 mA.
In order to disturb thermally said first portion and said
second portion, instead of applying a single arc, a plurality of electric arcs may
be applied in succession.
According to a further aspect, the present invention relates
to an optical amplifier comprising at least one optical amplification stage and
an optical filter as defined above, arranged in series with said optical amplification
According to a further aspect, the present invention relates
to an optical telecommunications system comprising at least one optical transmitter,
at least one optical receiver, an optical transmission line connecting said transmitter
to said receiver and at least one optical amplifier arranged along said transmission
line, in which said optical amplifier comprises at least one optical amplification
stage and an optical filter as defined above, arranged in series with said optical
Preferably, said optical amplifier comprises two optical
amplification stages and said optical filter is arranged between said two stages.
Alternatively, said optical filter is arranged downstream of said two stages.
According to a further aspect, the present invention relates
to an optical-fibre modal coupler comprising:
- an optical fibre which comprises a core and a cladding and through which an
optical signal can pass; and
- a coupling region formed in said optical fibre for producing a power transfer
between a first and a second propagation mode of said optical signal;
in which said optical fibre has, in cross section, an asymmetrical refractive index
profile in said coupling region.
According to a further aspect, the present invention relates
to a method for filtering an optical signal, said optical signal being transmitted
in a waveguide in the fundamental mode LP01, the method comprising the
- transmitting said signal through a first waveguide region having, in cross section,
an asymmetrical refractive index profile so as to transfer power from the fundamental
mode LP01 to the asymmetrical mode LP11;
- conveying said fundamental mode LP01 and said asymmetrical mode LP11
over a predefined distance so as to produce a relative phase shift depending on
said distance and the wavelength;
- transmitting said fundamental mode LP01 and said asymmetrical mode
LP11 through a second waveguide region having, in cross section, an asymmetrical
refractive index profile, so as to couple power between the fundamental mode LP01
and the asymmetrical mode LP11.
Further details may be obtained from the following description
which refers to the accompanying figures listed below:
- Figure 1 shows in schematic form a Mach-Zehnder filter of known type;
- Figure 2 shows the transmission spectrum of the filter according to Figure 1;
- Figure 3a shows a schematic and partial view of a Mach-Zehnder filter produced
in accordance with the invention;
- Figure 3b shows a cross section through line III-III of the filter of Figure
- Figure 3c shows an overall schematic view of the filter according to the invention;
- Figure 4 shows the refractive index profile of an optical fibre which can be
used in order to produce the filter according to Figure 3;
- Figure 5 shows in schematic form an apparatus for forming the modal coupling
regions of the filter according to the invention;
- Figure 6 shows in schematic form a step in the process for the production of
the filter according to Figure 3a, in which a predefined section of optical fibre
is struck by the electric arc of a fusion jointer;
- Figures 7a and 7b show respectively the refractive index profile of the fibre
of the filter according to the invention in a section thermally disturbed by the
process step according to Figure 6 and in a section not disturbed thermally;
- Figure 8 shows an apparatus which can be used for monitoring the coupling characteristics
of the filter during formation of the modal coupling regions;
- Figure 9 shows the transmission spectrum, obtained by means of the apparatus
according to Figure 8, of a filter produced in accordance with the invention;
- Figure 10 shows schematically an experimental apparatus for measuring the modal
coupling due to the asymmetrical variation in the refractive index profile of the
- Figures 11a and 11b show the results of a measurement carried out with the apparatus
according to Figure 10, following a regression (fitting) operation;
- Figure 12 shows a measurement of the losses of the filter according to the invention,
due to polarisation of the input signal;
- Figure 13 shows a different embodiment of the filter according to the invention;
- Figures 14a and 14b show by way of example the asymmetrical variations in the
index profile in different coupling regions of the filter according to Figure 13;
- Figure 15 shows the result of a measurement of the dependency of the filter
spectrum on the temperature;
- Figure 16 shows a diagram of a WDM optical transmission system; and
- Figure 17 shows an amplifier of the transmission system according to Figure
16, comprising the filter according to the invention.
With reference to Figure 3a, 1 denotes a fibre optical
filter of the Mach-Zehnder type. The filter 1 includes a dual-mode optical fibre
2 which has a length preferably of between 1 mm and 100 mm and comprises a core
3 and a cladding 4, both having the same longitudinal axis 5. The fibre 2 also has
a superficial protective coating 6 consisting of polymer material. The coating 6
is partially removed during the process for formation of the filter 1 (as shown
in the figure) and if necessary may be reapplied at a later stage.
The core 3 has a radius r1 and a refractive
index n1 and is composed of silica (SiO2) doped with one or
more elements which have the effect of raising the refractive index, such as, for
example, germanium (Ge) and phosphorus (P). As shown in Figure 3b, the cladding
4 comprises an inner region 4a and an outer region 4b, both of which are annular
in cross section. The inner region 4a borders with the core 3 (and therefore has
an internal radius equal to r1), has a refractive index n2
and has an external radius r2 equal to k·r1, where k
is a suitable coefficient, preferably between 2 and 6. Moreover, r1 is
preferably between 2.5 and 6.5 µm. The outer region 4b has a refractive index
n3 and an external radius r3 preferably equal to 62.5 µm.
Preferably, the difference n1-n2
between the refractive indices of the core 3 and the inner region 4a lies between
3.4·10-3 and 1.5·10-2. Moreover, the difference
n2-n3 between the refractive indices of the inner region 4a
and the outer region 4b is preferably between +1·10-3 and -2·10-3.
More preferably, n2 is substantially equal to n3 and the fibre
2 therefore has a refractive index profile substantially in the form of a step (step
index), as shown in Figure 4.
As is known, in the case of a fibre with a refractive step
index profile the cut-off wavelength &lgr;c is determined solely by
the radius r1 of the core 2 and the numerical aperture NA. In the present
case, the radius r1 and the numerical aperture NA are chosen so that
the fibre 2 is dual-mode in the spectral region currently of greatest interest for
optical telecommunications, i.e. between 1500 nm and 1650 nm.
Since the fibre 2 must be able to communicate with single-mode
fibres with known characteristics, it may be designed with a refractive step index
n1-n2 and with a radius r1 such as to have a distribution
of the fundamental mode substantially equivalent to that in the single-mode fibres
considered. In this way, power transfer of only the fundamental mode is ensured
between the fibre 2 and these single-mode fibres.
The inner region 4a has a composition such that it is thermo-refractive.
This composition comprises silica (Si), germanium (Ge), phosphorus (P) and fluorine
(F). The Applicant has ascertained that, with this composition, it is possible to
obtain, using the thermal disturbance technique described below, a variation in
the refractive index equal to at least 5·10-4. Advantageously, the
variation in the refractive index thus obtained may be greater than or equal to
1-10-3, even more advantageously greater than or equal to 2-10-3.
The Applicant has found that, in order to obtain an inner
region 4a with the abovementioned refractive index value and with the abovementioned
thermo-refractivity characteristics, the concentrations of the abovementioned dopants
in this region must lie within the following ranges:
- Ge: between 2% and 5%
- P: between 0.5% and 2%
- F: between 1% and 2%
Still with reference to Figure 3a, the fibre 2 has a first
and a second modal coupling regions 8, 9 positioned at a distance L from each other
along the axis 5. The modal coupling regions 8, 9 are formed by inducing thermally,
using the method described below, a variation in the refractive index &Dgr;n which
is asymmetrical in the inner region 4a. In practice, as shown in Figure 3b, a portion
7 (shown shaded in grey) of the inner region 4a, defining a substantially annular
sector in cross section, has a refractive index greater than that of the remainder
of the cladding cross section.
Owing to the presence, inside an optical fibre, of a zone
with a very asymmetrical refractive index profile, it is possible to achieve strong
power coupling between the fundamental mode LP01 and the asymmetrical
mode LP11. Each of the modal coupling regions 8, 9 therefore defines,
together with the fibre 2, a modal coupler. Preferably, the variation in the index
must be of a form (in the section considered) very similar to that of the mode LP11.
The section of fibre 2 lying between the two coupling regions
8, 9 - denoted by 10 - is referred to below as the "phase shift region" since it
defines the region in which the modes LP01 and LP11 undergo
a mutual phase shift &Dgr;&phgr; which is a function of the wavelength. The filter
1 therefore defines two coupling regions 8, 9 and a phase-shift region 10 lying
Furthermore, as shown in the schematic illustration in
Figure 3c, the filter 1 comprises a first and a second fibres 11, 12 of the standard
single-mode (SM) type, which are connected by means of respective joints at the
opposite ends of the fibre 2, so as to allow a substantially loss-free coupling
with the single-mode transmission fibres of the system in which the filter 1 is
placed. The fibres 11, 12 are single-mode in a spectral band lying between 1500
nm and 1650 nm. The filters 11, 12 define, respectively, an input for single-mode
signals Sin to be filtered and an output for the filtered single-mode
signals Sout. The fibres 11, 12 have geometric characteristics such that
they have a profile of the fundamental mode the same as that of the fibre 2, so
as to minimise the coupling losses therewith and perform modal filtering in order
to eliminate the mode LP11. The fibres 11, 12 preferably have a numerical
aperture NA of between 0.1 and 0.2 and an external radius (of the cladding) equal
to about 62.5 µm.
The operation of the filter 1 is described below. When
a single-mode optical input signal Sin reaches, via the first single-mode
fibre 11, the first coupling region 8, a transfer of power occurs from the mode
LP01 to the mode LP11 in a quantity dependent on the wavelength.
Subsequently, the modes LP01 and LP11 are propagated in the
phase shift region 10, at the end of which they have a phase difference &Dgr;&phgr;
expressed by the following relation:
where &Dgr;neff is the difference between the effective refractive
indices of the mode LP01 and the mode LP11 and &lgr; is
This phase difference is due to the different optical paths
followed by the modes LP01 and LP11 owing to their different
effective refractive indices neff and, therefore, to their different
speeds of propagation within the phase shift region 10. When the two modes LP01
and LP11 reach the second modal coupling region 9, they are combined
again, interfering constructively or destructively depending on the wavelength considered.
The outgoing signal from the second modal coupling region is further filtered upon
entering into the second single-mode fibre 12, with elimination of the mode LP11.
A single-mode signal Sout with a spectral form depending on the spectral
response of the filter 1 is therefore output from the fibre 12.
The process for manufacturing the filter 1 is described
The optical fibre 2 is made using the technique of modified
chemical vapour deposition (MCVD). In this process, in order to obtain the desired
composition in the thermo-refractive inner region 4a, then, in addition to the oxygen
and silicon tetrachloride (SiCl4) which are typically used in this process,
germanium tetrachloride (GeCl4), phosphorus oxychloride (POCl3)
and one of the following compounds of fluorine: Freon (CCl2F2),
sulphur hexafluoride (SF6) and silicon tetrafluoride (SiF4),
are also introduced into the deposition tube.
The coupling regions 8, 9 are then formed on the optical
fibre 2. The Applicant has found that the coupling regions 8, 9 may be formed by
applying, to the fibre 2, an asymmetrical thermal disturbance able to produce the
desired variation in the refractive index profile in the inner region 4a of the
cladding 4. The Applicant has also found that this thermal disturbance may be produced
by means of an electric arc.
With reference to Figure 5, 13 denotes an apparatus for
forming the coupling regions, comprising a fusion jointer 14 of known type, for
example a Fujikura model FSM-20CSII fusion jointer for optical fibres, and a fibre
moving device 15, able to perform micrometric displacements of the fibre 2 parallel
to its axis.
In order to induce an asymmetrical thermal disturbance,
the fibre 2 is positioned between the electrodes of the jointer 14, indicated by
16, 17, as shown in Figure 6.
The jointer 14 is then activated so as to produce an electric
arc 18 which causes sudden heating of the fibre 2 and, after the discharge, subsequent
rapid cooling thereof. Since the position of the fibre 2 is never perfectly symmetrical
with respect to the electrodes 16, 17, the electric arc 18 is usually formed only
on one side of the section of the fibre 2, as shown in the Figure. There is therefore
a temperature distribution inside the fibre 2 such as to cause an asymmetrical variation
in the refractive index &Dgr;n. This behaviour of the electric arc 18 may be observed,
for example, by positioning a videocamera (not shown) close to the electrodes 16,
17 of the jointer 14.
In order to form the other coupling region, the fibre 2
must be displaced parallel to its axis 4 by means of the fibre moving device 15
so as to arrange, between the electrodes 16, 17, a different portion of fibre 2,
the distance of which from the previously treated portion is exactly equal to L,
and apply again the electric arc to this portion.
The Applicant has manufactured, in order to carry out some
experimental measurements described below, a fibre 2 with the following characteristics:
- r1 equal to 4.7 µm;
- k (=r2/r1) equal to 4.3;
- n2 = n3;
- numerical aperture NA equal to 0.15;
- cut-off wavelength &lgr;c equal to 1630 nm;
- inner region 4a comprising (by scanning electron microscope (SEM) analysis):
95.2% silica (Si), 4% germanium (Ge), 0.8% phosphorus (P). The percentage of fluorine
(F), which cannot be determined using the SEM technique, was estimated at about
1.3% using the teaching of K. Abe, European Conference on Optical Fiber Communication,
Paris, 1996, Presentation II.4, taking into account that this concentration allows
to achieve the same refraction index value in the inner region 4a and in the outer
From this fibre, a filter 1 with the following additional
characteristics was produced:
- distance L between the coupling regions 8 and 9: 30 mm;
- numerical aperture of the fibres 11, 12: 0.12;
- cut-off length of the fibres 11, 12: 1200 nm;
- external radius of the cladding of the fibres 11, 12: 62.5 µm.
Figures 7a and 7b show, respectively, the refractive index
profile of the fibre thus obtained in the thermally disturbed section and in a section
which is not thermally disturbed. From Figure 7a it can be seen that the variation
&Dgr;n in the refractive index profile is asymmetrical in the section of the fibre
2 and has a maximum value of about 2.10-3.
The characteristics of the coupling regions 8, 9 are determined
by the power and the duration of the electric arc 18. The Applicant has noted that
it is not possible to establish precisely, on the basis of the parameters of the
electric arc 18, the amount of the variation in the index profile and, therefore,
the coupling factor. In order to verify the coupling properties of the regions 8,
9, it is possible to perform monitoring, during the writing process, of the extinction
ratio of the filter (which is correlated to the coupling factor), by means of a
spectral analysis. Figure 8 shows an apparatus 24 which can be used for monitoring
the coupling characteristics of the filter during the formation of the coupling
regions 8, 9. The apparatus 24 comprises a white light source 25 able to supply
wide-spectrum electromagnetic radiation to the fibre 2, a spectrum analyser 26 able
to analyse the spectrum of the light leaving the fibre 2, and a processing unit
27 connected to the spectrum analyser 26 for processing information supplied by
the said analyser.
Figure 9 shows the transmission spectrum, obtained by means
of the apparatus 24, of a filter 1 with the characteristics described above. From
this figure it is possible to note that the spectral form of a filter produced in
accordance with the invention is that typical of the interferential filters of the
Mach-Zehnder type, i.e. is periodic with a periodicity depending on the wavelength
considered. The extinction ratio thus obtained (namely the difference between the
minimum and maximum transmissivity of the filter expressed in dB) is equal to about
1.2 dB and the insertion losses are equal to about 0.4 dB. The Applicant has also
noted that, by optimising the process parameters, it is possible to obtain an extinction
ratio greater than 2.5 dB.
On the basis of the desired spectral response, the optical
filtering device according to the present invention may comprise, in a manner not
shown, several filtering stages arranged in cascade. In other words, this device
may comprise a plurality of filters 1 which are connected in series so as to have
a spectral response determined by the combination of the responses of the various
filters. As known from the text "Fiber Optic Networks", Prentice Hall, P.E. Green,
1993, page 123, in order to design a Mach-Zehnder filter with a desired spectral
behaviour, it is necessary to know the dispersion characteristic of the modes which
are propagated along the fibre, namely the value &Dgr;&bgr; (&lgr;) of the
difference between the propagation constants of the interfering modes. The modal
dispersion characteristic may be obtained by means of regression or "fitting" of
the spectral response (for example that shown in Figure 9) of a test filter of known
length. From this dispersion characteristic, it is possible, by means of digital
simulation, to determine the parameters of the interferometer, in particular the
distance L (or the distances L between the coupling points, in the case of several
interferometers arranged in cascade) and the values of the coupling coefficients,
which are required in order to produce the filter with the desired spectral response.
The efficiency with which the asymmetrical variation in
the index profile obtained using the technique according to the invention induces
coupling in the higher asymmetrical mode LP11 may be verified by means
of a suitable experimental test. For this purpose it is possible to use a measuring
apparatus such as that shown in Figure 10 and indicated therein by 19.
The measuring apparatus 19 comprises a laser source 20
able to supply to one of the ends of the fibre 2 a laser beam at the wavelength
of 1550 nm and a infrared videocamera 21 positioned so as to be able to detect the
light emitted from the fibre. In particular, the camera 21 is able to detect the
intensity profile of the electromagnetic field (known as "near field") emitted from
the fibre 2. The measuring apparatus 19 also comprises a processing unit 22 connected
to the camera 21 so as to receive from it a digital signal correlated with the optical
The intensity profile of the electromagnetic field detected
by the camera 21 is formed by the superimposition of the modes which are propagated
in the fibre 2 and, in mathematical terms, is defined by the square of the linear
combination of these modes. Each mode also has, associated with it, a multiplication
coefficient which determines its amplitude and, therefore, its weight within the
linear combination. In order to derive these coefficients it is possible to perform
a linear regression (or fitting) operation on the result of the experimental measurement.
In practice, based on the distribution of the fibre modes (LP01, LP11,
etc.), these modes are combined so as to obtain the intensity of the resultant field
which best approximates that measured.
The Applicant carried out a test using a fibre 2 having
the characteristics described above and provided with the coupling regions 8 and
9. Figures 11a and 11b show the linear regression (fitting) coefficients, the first
for the even modes (of the type LP0m) and the second for the odd modes
(of the type LP1m) obtained from the analysis of the fibre 2. These graphs
confirm that the only modes involved in the coupling are the modes LP01
and LP11. In the case in question, the values of the coefficients associated
with the modes LP01 and LP11 are equal to 0.79 and 0.21 respectively.
This measurement therefore confirms that the coupling induced by means of the asymmetrical
variation in the index profile produces a high modal selectivity, resulting in a
practically negligible contribution of modes other than the modes LP01
The Applicant has also noted that the coupling factor,
defined as being the ratio between the power transferred to the mode LP11
and that remaining in the mode LP01, increases with the intensity (in
other words with the amperage) of the electric arc. However, the Applicant has also
noted that if this intensity is too high, a geometric deformation of the fibre is
induced, in addition to a variation An in the refractive index in the fibre. This
deformation causes power losses which involve an increase in the insertion losses
of the filter and, therefore, a deterioration in the performance of the said filter.
It is therefore necessary to achieve a compromise between the desired coupling factor
(and therefore the desired extinction) and the resultant insertion losses. The Applicant
has ascertained that the electric arc must have a duration preferably of less than
400 ms, more preferably less than 300 ms, and a current intensity preferably between
8 and 14 mA, more preferably between 10 and 11 mA. More preferably, instead of a
single arc, a sequence of arcs with the abovementioned characteristics may be applied.
The Applicant also noted that, since the index profile
variation which causes coupling does not have a circular symmetry, the coupling
factor varies in accordance with polarisation of the light. The operation of the
filter 1 therefore depends on the polarisation of the incoming light. This dependency
is measured by evaluating, for each wavelength, the maximum variation which exists
in the attenuation spectrum of the filter with variation in the polarisation (PDL,
Polarisation Dependent Loss). Figure 12 shows the PDL measured, using a known technique,
on a filter 1 which has the characteristics described above. The mean value of the
measured PDL is about 0.4 dB, for a filter with an extinction ratio of about 2.6
The Applicant notes that this dependency on the polarisation
may be disadvantageous when the filter 1 is used in an amplification stage.
Figure 13 shows schematically a variation of the filter
according to the invention - denoted by 1' - able to reduce significantly the abovementioned
problem. The filter 1' differs from the filter 1 in that two further coupling regions
8' and 9' are present, preferably at a distance from each other equal to L. The
coupling regions 8' and 9' differ from the coupling regions 8, 9 in that the former
have an asymmetrical variation in the index profile which is perpendicular to that
of the latter. In particular the coupling regions 8' and 9' have, in cross section,
an annular sector 7' which is rotated through a right angle (90°) with respect
to the annular sector 7 of the coupling regions 8 and 9. Figures 14a and 14b show,
by way of example, the asymmetrical variations in the index profile in the coupling
regions 8 and 8', respectively (similar to those present in the regions 9 and 9',
respectively). The mutual distance between the coupling regions 8 and 8' and between
the coupling regions 9 and 9' is preferably the same, for example 100 µm. Since
this distance is very small, the undesirable effects of modulation of the signal
due to the presence of the additional coupling regions 8' and 9' is negligible.
As before, it is possible to produce an optical filtering
device comprising, in a manner not shown, a plurality of filters 1' connected in
The Applicant has also noted that the operation of the
filter 1 depends on the operating temperature. In particular, with a variation in
the temperature, the peaks in the spectral response of the filter 1 are displaced
in terms of wavelength. In order to verify the sensitivity to temperature of the
filter according to the present invention, a filter 1 with the characteristics indicated
above was positioned in a controlled-temperature chamber, in which the temperature
was varied (for example with a ramp-like variation) so as to cause the displacement,
in wavelength, of its resonance peaks. Figure 15 shows the results of this measurement.
In particular, the points measured and a regression (fitting) line for a filter
with a distance L of 20 mm are shown. It was found that, for each millimetre of
length of the filter, the position of the peak in the spectrum varies by about 0.0016
nm for each degree centigrade of variation in the temperature. The Applicant notes
that this dependency is substantially equivalent to that demonstrated by other interferential
filters of the Mach-Zehnder type.
The filter according to the present invention may be advantageously
used in a long-distance WDM (Wavelength Division Multiplexing) telecommunications
system, for example an undersea telecommunications system.
As shown in Figure 16, an optical telecommunications system
typically comprises a transmission station 32, a receiving station 33 and an optical
communications line 34 connecting the transmission station 32 and receiving station
33. The transmission station 32 comprises a plurality of optical transmitters 35,
each of which is able to transmit an optical signal at a respective wavelength.
Each optical transmitter 35 may, for example, comprise a source of the laser type
and a wavelength converter able to receive the signal generated by the laser and
transmit a signal at a predefined wavelength. A wavelength multiplexer 36 is connected
on its input side to the transmitters 35 so as to receive the plurality of signals
transmitted and has a single output connected to the communication line 34 in order
to transmit the wavelength multiplexed signals on the line. The transmission station
32 may also comprise an optical power amplifier 37, which is connected to the output
of the multiplexer 36, so as to impart to the signals transmitted the necessary
power for transmission along the line 34.
The receiving station 32 comprises a wavelength demultiplexer
38 connected at its input to the line 34 so as to receive the signals transmitted
and has a plurality of outputs into which the various wavelengths transmitted are
divided. The receiving station 32 also comprises a plurality of optical receivers
39, each connected to a respective output of the demultiplexer 38 in order to receive
a signal at a respective wavelength. Each receiver 39 may comprise a wavelength
converter to convert the wavelength of the signal into a wavelength suitable for
reception of the signal by a photo-detector connected optically to the said converter.
The receiving station 32 may also comprise a pre-amplifier 40 arranged upstream
of the demultiplexer 38 so as to impart to the signals transmitted the power necessary
for correct receiving thereof.
The communication line 34 comprises many sections of optical
fibre 41 (preferably single-mode optical fibre) and a plurality of line amplifiers
42 located at a distance from each other (for example a hundred kilometres or so)
and designed to amplify the signals to a power level suitable for transmission to
the next optical fibre section.
As shown schematically in Figure 17, at least one of the
amplifiers of the transmission system (i.e. the power amplifier 37, the pre-amplifier
and the line amplifiers 42), denoted here by 45, is a two-stage amplifier, i.e.
it comprises a first and a second active fibres 46, 47 for amplification of the
signals, connected in series. As shown, the filter 1 according to the invention
may be positioned between the two amplification stages so as to perform equalization
of the signals. Alternatively, the filter may be positioned downstream of the two
Lastly, the Applicant has found that the filter according
to the invention may be effectively used also as a temperature or deformation sensor
since its spectral response is sensitive to variations in temperature and length
in accordance with known laws. In particular, by detecting the displacement of predefined
points in the filter spectrum it is possible to determine the variation in the parameter
During operation as a temperature sensor, the sensor may
be used in order to measure the absolute temperature present in a given environment,
after being calibrated to a predefined temperature. In a similar manner, it may
be used to measure variations in temperature.
During operation as a deformation sensor, the filter 1
is applied to a body liable to undergo deformation. The variation in the spectral
response of the filter 1 following deformation of the body provides a measurement
of the said deformation.