U.S. patent application number 09/848296 was filed with the patent office on 2001-12-13 for formation of a refractive index grating.
This patent application is currently assigned to British Telecommunications Public Limited Company. Invention is credited to Kashyap, Raman.
Application Number | 20010051020 09/848296 |
Document ID | / |
Family ID | 26140494 |
Filed Date | 2001-12-13 |
United States Patent
Application |
20010051020 |
Kind Code |
A1 |
Kashyap, Raman |
December 13, 2001 |
Formation of a refractive index grating
Abstract
An apodised refractive index grating is recorded in a
photosensitive optical fibre by forming first and second component
interference patterns with different pitches, that are recorded in
the grating such as to result in apodisation. The component
patterns (16, 17) are spatially in phase in a central region and
move progressively out of phase towards the ends of the patterns.
The patterns may be recorded sequentially or concurrently. The
fibre may be stretched one or cyclically.
Inventors: |
Kashyap, Raman; (Suffolk,
GB) |
Correspondence
Address: |
Nixon & Vanderhye P.C.
8th Floor
1100 North Glebe Rd.
Arlington
VA
22201-4714
US
|
Assignee: |
British Telecommunications Public
Limited Company
|
Family ID: |
26140494 |
Appl. No.: |
09/848296 |
Filed: |
May 4, 2001 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
09848296 |
May 4, 2001 |
|
|
|
09011995 |
Mar 9, 1998 |
|
|
|
09011995 |
Mar 9, 1998 |
|
|
|
PCT/GB96/03079 |
Dec 12, 1996 |
|
|
|
Current U.S.
Class: |
385/37 ; 359/11;
359/566; 430/290 |
Current CPC
Class: |
G02B 6/02133 20130101;
G02B 6/02152 20130101; G02B 6/02138 20130101; G02B 6/02085
20130101 |
Class at
Publication: |
385/37 ; 359/566;
359/11; 430/290 |
International
Class: |
G02B 006/34; G02B
005/18 |
Foreign Application Data
Date |
Code |
Application Number |
Dec 12, 1995 |
EP |
95309031.3 |
Mar 29, 1996 |
GB |
9606781.4 |
Claims
1. A method of recording an apodised refractive index grating in a
photosensitive optical medium with a pattern of optical radiation,
that comprises producing a plurality of spatially periodic
component optical patterns for recording a sequence of elements
that form the grating, with a relative spatial phase which varies
along the sequence in such a manner as to result in apodisation of
the grating recorded in the optical medium.
2. A method according to claim 1 wherein the effective refractive
index (n.sub.eff) of the optical medium is substantially constant
along the recorded grating.
3. A method according to claim 1 or 2 wherein said relative phase
progressively changes in directions away from an intermediate
region of the component patterns towards the ends thereof.
4. A method according to claim 3 wherein the component patterns
have zero relative spatial phase in the intermediate region.
5. A method according to claim 4 wherein the component patterns
have a relative phase of .+-..pi./2 in respect to the spatial
periodicity of the patterns, at opposite ends thereof.
6. A method according to claim 3, 4, or 5, wherein the intermediate
region is disposed centrally of the component patterns.
7. A method according to any preceding claim wherein the component
patterns are formed sequentially.
8. A method according to any preceding claim wherein the component
patterns are optical interference patterns.
9. A method according to claim 8 including causing beams of optical
radiation to interfere to produce a first of the component
interference patterns, and thereafter introducing a phase shift
across the width of at least one of said beams, so as to form a
second of the component interference patterns.
10. A method according to claim 8 including introducing a wedge of
optically transparent material into one of said beams so as to
produce said phase shift for said second component interference
pattern.
11. A method according to claims 1 to 10 including forming an
interference pattern, arranging the optical medium in a first
disposition relative to said interference pattern so as to provide
the first component interference pattern to be recorded in the
medium, and thereafter arranging the medium in a second disposition
relative to the interference pattern so as to provide the a second
component interference pattern to be recorded in the medium.
12. A method according to claim 11 including rotating the optical
medium relative to the interference pattern between the first and
second dispositions.
13. A method according to claim 11 including altering the length of
the optical medium in order to achieve said first and second
dispositions.
14. A method according to claim 13 including stretching the
medium.
15. A method according to claim 13 or 14 including stretching the
optical medium cyclically
16. A method according to claim 13, 14 or 15 including stretching
the medium from one end only of the recorded pattern.
17. A method according to claim 13 including compressing the
medium.
18. A method according to 13, 14, 15 or 16 including repeatedly
recording the pattern along the medium
19. A method according to any one of claim 13 to 18 wherein the
pattern is provided from a phase mask.
20. A method according to claim 19 including altering the length of
the phase mask so as to produce the component patterns.
21. A method according to claim 19 wherein the optical medium is
stretched symmetrically about the centre of the resultant
grating.
22. A method according to claim 19 wherein a beam of said radiation
is scanned along the phase-mask and the optical medium is
cyclically stretched such that the condition: 4 f > v w is met,
where f is the frequency of stretching, v is the scanning speed of
the beam and w is the diameter of the beam spot.
23. A method according to any one of claims 1 to 6 wherein the
component patterns are formed concurrently.
24. A method according to claim 23 wherein beams of optical
radiation are caused to interfere to produce said component
patterns, said beams having a predetermined spectral content
whereby light at different wavelengths interferes to produce said
patterns.
25. A method according to any one of claims 1 to 6 wherein a phase
mask that includes first and second mask patterns, is placed
adjacent to the waveguide and light is directed through the phase
mask so as to produce the component patterns to be recorded in the
optical medium in the waveguide.
26. A method according to claim 25 including selectively
illuminating the mask patterns in the phase mask sequentially.
27. A method according to any preceding claim including recording
the refractive index grating in an optical waveguide that is
photosensitive to the optical radiation.
28. A method according to claim 27 wherein the waveguide comprises
an optical fibre.
29. A waveguide including an apodised refractive index grating
formed by a method according to any preceding claim.
30. An optical chirp filter comprising a grating which is a
composite of a first component grating and a second component
grating, the phase difference between the first and second
component gratings increasing from the centre of the said grating
to produce apodisation thereof.
31. An optical wavelength division multiplexed communications
system comprising a transmitting station and a receiving station
coupled by an optical waveguide, the receiving station including a
plurality of chirp filters according to claim 27, each filter being
arranged to compensate for dispersion of an optical signal in a
different WDM channel received from the waveguide.
Description
FIELD OF THE INVENTION
[0001] This invention relates to a method of recording an apodised
refractive index grating in a photosensitive optical medium and has
particular but not exclusive application to forming gratings in
optical fibres.
BACKGROUND
[0002] It is known that the refractive index of an optical fibre
can be altered by exposing it to high intensity light. Germanium
doped fibre exhibits a photosensitivity in this manner and the
effect can be used to form a so-called refractive index grating in
the fibre. Reference is directed K. O. Hill et al,
"Photosensitivity in Optical Waveguides: Application to reflection
filter fabrication", Appl. Phys. Lett., Vol 32, no. 10, 647 (1978).
The grating can be produced by forming an optical interference
pattern with two interfering beams, and exposing the optical fibre
to the interference pattern, so as to record a grating in the
fibre.
[0003] The interference pattern may be formed by directing an
optical beam longitudinally through the fibre and reflecting it
back along its path through the fibre, so as to produce a standing
wave pattern, which becomes recorded in the fibre due to its
photosensitivity. In an alternative method, beams derived from a
coherent source are directed transversely of the length of the
fibre, so as to interfere with one another and produce an
interference pattern externally of the fibre, which becomes
recorded in the fibre as a result of its photosensitivity. A block
for producing an external interference pattern for this purpose is
described in EP-A-0523084.
[0004] Another way of forming the grating is to use a phase mask in
which the desired amplitude pattern has been recorded
holographically as a mask pattern. The phase mask is placed
adjacent to the fibre and the illuminated laser light, so as to
expose the fibre to the holographic pattern. Reference is directed
to K. O. Hill et al "Bragg grating fabricated in monomode
photosensitive fiber by u.v. exposure through a phase mask" Appl.
Phys. Lett. Vol. 62, No. 10, 1035 (1993).
[0005] For a general review of refractive index gratings, reference
is directed to "Photosensitive Optical Fibres: Devices and
Applications" R. Kashyap, Optical Fiber Technology 1, 17-34
(1994).
[0006] Also, reference is directed to U.S. Pat. No. 4,474,427 to
Hill and PCT/GB91/01968 (WO92/08999) which disclose the formation
of more than one refractive index grating pattern in a common
optical fibre.
[0007] Refractive index gratings, which operate as Bragg gratings,
have many applications in optical data communication systems as
discussed by Kashyap, supra and in particular can be used as
wavelength filters. It is well known that the large bandwidth
offered by an optical fibre can be used to transmit data at a
number of different wavelengths, for example by wavelength division
multiplexing (WDM). It has been proposed to use refractive index
gratings to separate information from adjacent WDM channels.
Conventionally, optical telecommunication networks transmit data in
channels centred on 1.3 .mu.m and 1.5 .mu.m. In either of these
wavelength regions, a Bragg grating can be used to reflect out a
narrow wavelength channel of the order of .mu.m or less, in order
to permit WDM demultiplexing. A series of gratings can be provided
to select individual closely spaced channels. The gratings exhibit
a main wavelength peak centred on the wavelength of the channel to
be filtered, but each grating also exhibits a series of side lobes
at harmonics of the wavelength peak, which produce reflection in
adjacent channels, resulting in cross-talk. As a result, it has
proved necessary to apodise the Bragg gratings so as to suppress
the effect of the side lobes and reduce the cross-talk.
[0008] Prior apodisation techniques will now be discussed.
Referring to FIG. 1, this shows a conventional method of forming a
refractive index grating in an optical fibre, in which light from a
laser source 1 is fed through a beam spitter 2 in order to form
coherent beams 3, 4, which are directed by a mirror arrangement 5,
6 so as to interfere with one another in region 7 adjacent to an
optical fibre 8 which exhibits photosensitivity at the wavelength
of operation of the laser 1. The result is an optical interference
pattern, which is recorded in the fibre as a result of its
photosensitivity. The result of the recording is shown in FIG. 2.
The spatially periodic intensity of the interference pattern
produces a corresponding pattern of refractive index variations
along the length of the fibre, which in FIG. 2 are schematically
shown as refractive index regions n.sub.1 and n.sub.2. These
regions act as a reflection grating in a manner well known per se.
The grating has a wavelength dependent reflection characteristic
with a main lobe centred at a particular wavelength depending upon
the periodic spacing of the refractive index regions n.sub.1,
n.sub.2, together with is a series of side lobes at harmonics of
the centre wavelength. The reflection wavelength .lambda..sub.Bragg
is given by
.lambda..sub.Bragg=2.LAMBDA.n.sub.eff/N
[0009] where .LAMBDA. is the period of diffraction pattern and
n.sub.eff is the effective refractive index of the waveguide. N is
an integer.
[0010] Referring to FIG. 2b which shows the variation in refractive
index recorded in the fibre, the spatially periodic function has an
envelope 10 which in the simple example shown in FIG. 2b is
theoretically flat for an infinitely long grating. This is shown
again in FIG. 3a, with the periodic function omitted. The
corresponding spectral characteristic for the grating, i.e. the
response in the wavelength domain, is shown in FIG. 3b and it can
be seen that the grating exhibits a main lobe 11 and a series of
side lobes 12.sub.n, 13.sub.n on either side of the main lobe. When
the grating is used as an optical filter e.g. in a WDM
demultiplexer, the spacing of the grating pattern is chosen so that
the main lobe 11 corresponds to the centre wavelength of the WDM
channel, but a problem arises in that the side lobes 12, 13 extend
into adjacent wavelength channels for the WDM system, particularly
when the channels are closely spaced in wavelength. The side lobes
thus will produce reflection in the adjacent channels and result in
cross-talk.
[0011] Apodisation suppresses the effect of the side lobes. This
has been achieved hitherto in a number of different ways. Referring
to FIG. 1, the grating pattern formed in the region 7 will not in
fact have a constant amplitude along its length and as a result,
the refractive index pattern recorded in the fibre does not in
practice have a flat envelope 10 as shown in FIG. 2b. Actually, the
beams 3, 4 have an approximately Gaussian amplitude spread across
their physical width, with the result that the envelope 10 in
practice has a shape more like that shown in FIG. 4a. It can be
shown that suppression of the side lobes will be achieved if the
envelope 10 has a shape which tapers from a central region towards
its opposite ends, for example in accordance with the function
cos.sup.2z along the length z of the recorded grating. In the past,
this has been attempted by modifying the amplitude distribution
across the width of the beams 3, 4. The corresponding spectral
response of the filter is shown in FIG. 4b, from which it can be
seen that the effect of side lobes is suppressed.
[0012] For gratings recorded in a phase mask, apodisation has been
achieved by varying the intensity of the pattern across the mask,
or by selective destruction of the phase pattern recorded in the
mask. Reference is directed to "Apodised in-fibre Bragg grating
reflectors photoimprinted using a phase mask", B. Malo et al
Electronics Letters Feb. 2, 1995, Vol 31, No. 3, pp 223-225; and
also to "Apodisation of the spectral response of fibre Bragg
gratings using a phase mask with variable diffraction efficiency",
J. Albert et al, Electronics Letters, Feb. 2, 1995, Vol 31, No. 3
pp 222-223.
[0013] However, a problem with all of these prior techniques is
that the side lobes are not suppressed completely, due to the fact
that the overall refractive index exhibited by the fibre n.sub.eff
varies along the length z of the grating. It will be recalled that
the value of refractive index n recorded in the fibre is a function
of the intensity of the illuminating light, so that with the
configuration shown in FIG. 4a, the effective refractive index
n.sub.eff varies along the length of the grating in a non-uniform
manner 14. This non-uniform variation itself produces chirp in the
Bragg wavelength of the grating, and as a result side lobes in the
spectral response of the structure.
[0014] Hitherto, post-processing techniques have been used in order
to linearize n.sub.eff. However, these techniques have been
difficult to implement in practice. Reference is directed to Hill
et al, supra.
[0015] An alternative apodisation technique has recently been
proposed in "Moving fibre/phase mask scanning beam technique for
enhanced flexibility in producing fibre gratings with uniform phase
mask" M. J. Cole et al, Electronics Letters, Aug. 17, 1995, Vol 31,
No. 17, pp 1488-1490. In this technique, the grating is recorded in
a manner generally shown in FIG. 1 and, additionally, a
piezoelectric device is moved along the fibre from a central
position in the grating, during its formation, so as to apply
vibration to the fibre, the amplitude of which increases towards
the exterior edges. In this way, the recorded pattern is "blurred"
towards the ends of the recorded grating which has the effect of
apodising the grating, but without reducing the intensity of the
recording light towards the ends of the grating as in the
previously described methods, with the result that n.sub.eff need
not vary significantly along the length of the grating.
[0016] A method of providing a surface relief diffraction grating
for use in a distributed feedback (DFB) optical fibre laser is
described in GB-A-2209408. The grating is formed by exposing a
layer of photoresist on an optical fibre to two different optical
interference patterns of different periodicities, produced by
interfering beams of optical radiation. The resulting, exposed
composite pattern formed in the photoresist is then developed, and
the fibre is etched using the developed pattern as a mask, to
provide a surface relief pattern in the fibre. The two component
patterns are chosen so as to support a common longitudinal mode in
the output of the laser. However, the configuration does not
produce apodisation because the surface grating pattern has an
effective refractive index which varies along the length of the
optical fibre, which results in unwanted sidelobes in the
wavelength characteristic.
[0017] The present invention provides a technique for controlling
the spectral characteristic of a refractive index grating recorded
in a photosensitive optical medium, which can be used to produce
apodisation.
[0018] The present invention provides a method of recording an
apodised refractive index grating in a photosensitive optical
medium with a pattern of optical radiation, comprising producing a
plurality of spatially periodic component optical patterns for
forming the grating, with a relative spatial phase which varies
along the length thereof in such a manner as to result in
apodisation of the grating recorded in the optical medium.
[0019] In accordance with the invention, the effective refractive
index of the optical medium may be substantially constant along the
length of the recorded grating, so as to provide effective
apodisation.
[0020] The relative phase of the component patterns may
progressively increase in directions along the patterns away from
an intermediate region towards ends thereof. The patterns may have
zero relative phase in the intermediate region, and a relative
phase of .+-..pi./2 with respect to the spatial periodity of the
pattern, at the ends thereof.
[0021] The overlying component patterns may be formed sequentially.
They may comprise optical interference patterns. The interference
patterns may be formed by causing beams of optical radiation to
interfere to produce a first of the patterns, and thereafter
introducing a phase shift across the width of at least one of the
beams, so as to form a second of the interference patterns. An
optically transparent wedge may be used to introduce the phase
shift.
[0022] Alternatively, the optical medium may be moved relative to
the pattern, so as to provide the first and second component
patterns. The waveguide in one example is rotated through a small
angle, and in another, is stretched between recording the patterns
in the optical medium.
[0023] The component optical patterns may be formed concurrently.
For example beams of optical radiation with a predetermined
spectral content may be caused to interfere so that light at
different wavelengths interferes to produce the component patterns
concurrently.
[0024] In another method, the component patterns are derived from
corresponding patterns recorded in a phase mask.
[0025] The invention has particular application to recording an
apodised grating in an optical waveguide such as photosensitive
optical fibre e.g. a germanium doped fibre that is photosensitive
to u.v. radiation. The recording method according to the invention
has the advantage that the recorded pattern is both apodised and
has an average intensity which need not vary significantly along
the length of the pattern, so that the average refractive index
n.sub.eff need not vary, thereby avoiding chirp and resultant cross
talk.
BRIEF DESCRIPTION OF THE DRAWINGS
[0026] In order that the invention may be more fully understood,
embodiments thereof will now be described, reference being had to
the accompanying drawings in which:
[0027] FIG. 1 illustrates a prior art method of forming a
refractive index grating in an optical fibre;
[0028] FIG. 2a illustrates the refractive index pattern of the
grating formed in the fibre;
[0029] FIG. 2b is a graph of the refractive index variation along
the length of the fibre;
[0030] FIG. 3a illustrates the envelope of the refractive index
variation along the length of the fibre, and;
[0031] FIG. 3b illustrates the corresponding spectral
characteristic of the grating;
[0032] FIGS. 4a and 4b correspond to FIGS. 3a and 3b, but for an
amplitude mask apodised grating;
[0033] FIG. 5 is a schematic illustration of apparatus for
recording a grating by a first method according to the
invention;
[0034] FIG. 6 illustrates apparatus for recording a grating by a
second method in accordance with the invention;
[0035] FIG. 7 is a schematic illustration of the recording of a
grating by a third method in accordance with the invention;
[0036] FIG. 8 illustrates a method of recording a grating by a
fourth method according to the invention;
[0037] FIG. 9 is a graph of the wavelength distribution of the
laser used in the method described with reference to FIG. 8;
[0038] FIG. 10 illustrates apparatus for performing a fifth
recording method in accordance with the invention;
[0039] FIG. 11 illustrates first and second component interference
patterns for recording the grating;
[0040] FIG. 12 is a schematic illustration of the combination of
the intensities of the first and second component interference
patterns;
[0041] FIG. 13 is an enlarged view of the pattern shown in FIG. 12,
adjacent the end 20a;
[0042] FIG. 14 is an enlarged view of the pattern shown in FIG. 12,
for the central region 19;
[0043] FIG. 15 is a graph of the wavelength response of the
apodised grating recorded by the method of the invention, in
reflection along the optical fibre;
[0044] FIG. 16 illustrates an apparatus for performing a sixth
method according to the present invention;
[0045] FIG. 17 is a graph of the reflectivity of a grating formed
according to the present invention with the actual and predicted
reflectivity of a prior art grating for comparison;
[0046] FIG. 18 is a graph showing the delay of a grating formed
according to the present invention with the actual and predicted
reflectivity of a prior art grating for comparison; and
[0047] FIG. 19 is a schematic block diagram of a communications
system employing gratings made according to the present
invention.
DETAILED DESCRIPTION
[0048] A first example of a grating formation method in accordance
with the invention will now described with reference to FIG. 5. The
method can be considered as a modification of the method described
with reference to FIG. 1, in which like parts are marked with the
same reference numbers. The two interfering beams 3, 4, which, in
FIG. 5 are shown to have a given width, interfere in region 7 to
form an interference pattern which is recorded in the optical fibre
8 in the manner generally as previously described. The laser 1,
beam splitter 2 and mirrors 5, 6 have been omitted from FIG. 5 in
order to simplify the drawing.
[0049] In accordance with the invention, first and second component
interference patterns are produced and individually recorded in the
fibre so as to produce the apodisation of the grating. The
component grating patterns have slightly different spatial
periodicities, chosen so that their combined effect is to suppress
the side lobes in the wavelength response of the recorded grating,
as will be explained more fully hereinafter.
[0050] In FIG. 5, the first component interference pattern is
recorded in the fibre by producing interference between beam 4 and
beam 3 shown in dotted outline. Thereafter, a transparent wedge 15
placed in the beam 3 in order to introduce a progressive phase
shift in the wavefront of the beam across its width. This
interferes with the beam 4 to produce a second component
interference pattern. There is a progressive small increase in
spacing between the successive peaks and troughs of the second
interference pattern in comparison with the first pattern and this
will be explained in more detail hereinafter with reference to FIG.
11. Thus, the second, slightly different pattern is recorded in the
fibre 8, overlying the first pattern formed by the beams 3a, 4. In
order to form the second component pattern, the beam 3 may need to
be shifted to the position 3b shown in FIG. 5 in order to produce
the correct overlying alignment of the first and second component
patterns.
[0051] Referring now to FIG. 11, this shows the amplitude of the
first and second component interference patterns 16, 17 that are
produced. The first pattern 16 is shown in dotted outline whereas
the second pattern 17 is shown as a solid line. FIG. 11 shows the
component patterns at three positions across the width of the
interference pattern 7, namely at the left side 7.sub.l the centre
7.sub.c and the right side 7.sub.r. The second component pattern 17
has a slightly different periodicity from the pattern 16 due to the
phase shift introduced by the wedge 15. In the central region
7.sub.c, the patterns overlie one another, but due to their
different periodicities, they become progressively out of phase
towards the side edges of the interference pattern so that in
positions 7.sub.l and 7.sub.r the patterns are out of registry, as
shown in FIG. 11.
[0052] It will be understood that the combination of the
intensities corresponding to the two amplitude patterns 16, 17
shown in FIG. 11 is recorded in the photosensitive waveguide 8
(FIG. 5) as variations in the refractive index of the waveguide.
The resultant combined intensity pattern produced by the component
amplitude patterns 16, 17 is shown in FIG. 12 and consists of a
spatially periodic function having an envelope 18. The value of the
cyclic spatial intensity variations for the function shown in FIG.
12 is highest in a central region 19 of the function and
progressively decreases towards the opposite ends 20a, 20b in the
direction z along the waveguide. A more detailed view of the
intensity function of FIG. 12 adjacent the end 20a is shown in FIG.
13 and a more detailed view of the region 19 is shown in FIG. 14.
The refractive index grating produced by this intensity pattern is
thus apodised as a result of the shape of the envelope 18.
Furthermore, the shape of the envelope 18 has the advantage that
the average intensity remains constant along the length z so that
n.sub.eff is substantially constant along the length of the
grating, which minimises side lobes in the grating characteristics
produced by chirp that was described previously with reference to
FIG. 4A.
[0053] A more detailed discussion of the function shown in FIGS. 12
to 14 will now be given. The spatially periodic amplitude of the
first and second component interference patterns 16, 17 in the
direction z will be referred to as A.sub.1(z) and A.sub.2(z), and
the resultant intensity pattern I(z) shown in FIG. 12 can be
written as the sum of the squares of the intensity patterns, i.e. 1
I ( z ) = K 2 ( A 1 2 + A 2 2 ) ( 1 )
[0054] where K is a constant.
[0055] Each of amplitudes A.sub.1(z) and A.sub.2(z) can be
represented as a spatial cosine function, i.e.
A.about.cos .beta.z
[0056] where the period .LAMBDA. of the diffraction grating pattern
is given by:
.LAMBDA.=2.pi./.beta.
Thus A.sub.1=p cos .beta..sub.1 z (2)
and A.sub.2=q cos .beta..sub.2z (3)
[0057] where p and q are constants.
[0058] From equations (1), (2) and (3), it follows that
I(z)=K/2 (p.sup.2 cos .sup.2 .beta..sub.1 z+q.sup.2 cos .sup.2
.beta..sub.2z)
[0059] where K is a constant
[0060] which can be written as
I(z)=K/2 (P.sup.2 cos .sup.2 .beta..sub.1 z+Q.sup.2 cos .sup.2
.beta..sub.2z) (4)
[0061] where P and Q are constants. For the purpose of simplicity
in the following analysis, P and Q are assumed to be of value
1.
[0062] The form of equation (4) will now be considered in detail by
way of example at the end of point 20a and in the central region 19
in order to explain the refractive index variations that are
recorded in the waveguide as a grating pattern.
[0063] In the central region 19, the two component patterns 16 and
17 are substantially in phase, with the same spatial periodicity,
so that in region 19 .beta..sub.1=.beta..sub.2=.beta.. Thus,
equation (4) reduces to
I.sub.(19)=1/2(cos .sup.2 .beta.z+cos .sup.2 .beta.z)
i.e. I.sub.(19)=cos.sup.2 .beta.z
[0064] It can be shown that over one spatial period of the pattern
i.e. .beta.=0.fwdarw.2.pi., the inensity I.sub.(19) has an average
value <I.sub.(19)>=1/2 in the arbitrary units of this
analysis.
[0065] Considering now the end point 20a of the envelope 18, the
patterns 16, 17 are arranged to become progressively out of phase
from the in-phase condition in region 20, so that at the opposite
ends of the envelope, e.g. at end point 20a, the patterns are
90.degree. out of phase, i.e.
.beta..sub.1=.beta..sub.2+.pi./2=.beta.. Thus, at the end point
20a, equation (4) reduces to: 2 I 20 = 1 / 2 ( cos 2 2 + cos 2 ( z
+ / 2 ) ) = 1 / 2 ( cos 2 z + sin 2 z ) = 1 / 2 ( in the arbitrary
units of this analysis )
[0066] It will be therefore understood from the foregoing, and from
an inspection of FIG. 12, that the average intensity is constant
along the z dimension of the envelope 18, shown as line 21. Thus
n.sub.eff as recorded in the fibre 8, is constant along its length,
which avoids chirp that would arise if n.sub.eff were to vary.
[0067] In an example of a grating recorded in a germanium doped
photosensitive fibre 8, u.v. light from a laser operating
c.w..about.100 mw at a wavelength of 244 nm produced the first
interference pattern 16 in region 7 of FIG. 5, of a length z=4-6 mm
and a transverse dimension of 40 .mu.m. The spatial period of the
interference pattern was of the order of 1 .mu.m. Then, the wedge
15 was inserted into beam 3 in order to produce the second pattern
17. The wedge was made of SiO.sub.2 with a refractive index n=1.46,
and a wedge angle of 5" of arc. The second pattern 17 was spatially
in phase with the first pattern in the central region 19 and the
relative spatial phase of the patterns progressively increased
outwardly from the central region 19 towards the ends 20a, 20b
where the phase difference was .pi./2. The optical fibre consisted
of a silica fibre with an outside diameter of 125 .mu.m, codoped
with Ge/B to provide a core of 4 .mu.m diameter. The exposure time
for each component pattern was approximately ten minutes.
[0068] The resultant spectral characteristic of the apodised
grating recorded in the fibre 8 is shown in FIG. 15, as trace 23.
For comparison purposes, the wavelength characteristic for a
grating produced by only one of the patterns 16 or 17 is shown as
trace 22, from which the suppression of the side lobes produced by
the apodisation can be clearly seen. The characteristics shown in
FIG. 15 were determined by launching relatively broadband laser
radiation along the core of the fibre 8 and measuring the spectral
response of the radiation reflected by grating, using conventional
techniques.
[0069] Referring now to FIG. 6, the second method of recording an
apodised grating in accordance with the invention will now be
described. This can be considered as a modification of the method
described with reference to FIG. 5. In the method of FIG. 6, the
beams 3, 4 produce interference in region 7, in the manner
described previously. The first interference pattern is recorded in
the fibre 8 when it is in position A so that the fibre is in
position 8.sub.1. Thereafter, the fibre is moved through a small
angle e.g. .about.3.degree. for a pattern of length z=4 mm, and the
second pattern is recorded, whilst the fibre is in position B,
lying along line 8.sub.2 shown in dotted outline. The plane in
which the fibre is rotated may lie within the beams 3, 4 or may be
transverse to the beams, for example in a horizontal plane for the
configuration shown in FIG. 6.
[0070] Thus, by the method described with reference to FIG. 6, the
first and second component interference patterns are formed in
relation to the fibre 8, which are recorded therein so as to
produce a combined pattern in which the fibre grating is
apodised.
[0071] A third method in accordance with the invention will now be
described with reference to FIG. 7. In this example, the beams 3
and 4 are directed to the fibre 8 as before, and the fibre is
subjected to different levels of longitudinal stress. Piezoelectric
devices 22, 23 are attached to opposite ends of the fibre 8. The
beams 3, 4 produce an interference pattern in region 7.
[0072] Firstly, the fibre is subject to a first relatively low
level of stress, during which the piezoelectric devices 22, 23 are
unenergized. A first component grating pattern is recorded in the
fibre during this period. Thereafter, the devices 22, 23 are
energized so that the fibre is stretched by a small amount
corresponding to a period A of the interference pattern. The
interference pattern formed in region 7 is then recorded again as a
second component pattern in the fibre 8, with the level of
stretching being maintained during exposure for the second pattern.
Thereafter, when the exposure is completed, the fibre is released
from the piezoelectric devices 22, 23. When the stretching is
released, the spatial periodicity of the second pattern becomes
slightly compressed as a result of the release of the fibre stress,
so that the second pattern has a slightly smaller periodicity than
the first pattern. The patterns are arranged so that they are
spatially in phase in their central regions, and are b 90.degree.
out of phase at the opposite ends, so the resulting combination of
the first and second patterns recorded during the first and second
exposures of the fibre, produces an apodised grating.
[0073] In a modification the piezo electric devices 22, 23 are
driven by an oscillator (not shown) e.g. at a frequency of about 5
Hz, during the exposure, which may take several minutes. This
results in the desired apodised pattern.
[0074] In the examples of the method according to the invention
described so far, the first and second component patterns have been
recorded sequentially. However, it is possible to achieve
simultaneous recording of the patterns and an example will now be
described with reference to FIG. 8.
[0075] Light from a laser source 24 is directed through a phase
mask 25, which acts as a beam splitter, so as to form two phase
coherent beams 26, 27 which pass through respective reflective
corner cubes 28, 29 so as to be reflected back on paths 30, 31 to
mirrors 32, 33. The mirrors are adjusted so as to reflect the beams
along paths 34, 35 which converge at an angle .theta. upon the
photosensitive optical fibre 8. The spectral content of the output
of the laser 24 is shown schematically in FIG. 9 and consists of a
narrow Gaussian distribution of wavelengths with a peak wavelength
.lambda..sub.max. The two beams 34, 35 interfere and produce an
interference pattern in the region 7 of the fibre. The interference
pattern can be considered as a superposition of patterns produced
at each of the component wavelengths that makes up the distribution
shown in FIG. 9. It can be shown that the resulting superposition
gives rise to an apodised grating pattern recorded in the
fibre.
[0076] A fifth example of a method according to the invention will
now be described with reference to FIG. 10. In this example, two
holographic phase mask patterns are used to record the first and
second component patterns in the fibre 8. The phase mask patterns
may be formed one overlying the other in the same phase mask, and,
in FIG. 10, the first and second component patterns are shown as
patterns P1 and P2 formed in a phase mask 36. An optical system 37
shown schematically, is operable to focus a beam of light either
onto the pattern P1 or the pattern P2 and cast a corresponding
holographic reference pattern onto optical fibre 8. The optical
system 37 thus, records interference patterns derived
holographically from the patterns P1, P2 sequentially in the fibre
and the result of the two component patterns, when recorded, is to
produce a grating which exhibits apodisation.
[0077] In a modification, the patterns P1 and P2 may be recorded
side by side in the mask 36, and the mask is moved between
exposures to align the patterns with the optical source 37 and the
fibre 8. In another modification, a single phase mask pattern P1 is
used and the fibre 8 or the phase mask 36 is stretched to produce
the second component pattern to be recorded in the fibre. Also
instead of stretching, the phase mask can be compressed to alter
the periodicity of the pattern P1 so as to provide the second
component pattern to be recorded in the optical fibre. The
compression technique can also be applied to the optical medium.
Although an optical fibre cannot easily be longitudinally
compressed, the compression technique is particularly useful for
recording apodised gratings in planar waveguides, which cannot
readily be stretched but can be longitudinally compressed.
[0078] A sixth example of the invention will now be described with
reference to FIG. 16 in which the fibre 8 is placed immediately
behind phase mask 36 and held in chucks in assemblies 39, 40 that
include piezoelectric actuators that correspond to the actuators
22, 23 shown in FIG. 7. The phase mask 36 is formed from silica,
etched by a standard e-beam technique, with a 100 mm long
step-chirped grating with a chirp of 0.75 nm. The grating comprises
200 sections, each 0.5 mm in length, mimicking a near continuous
chirp. A beam 41 of UV radiation at 244 nm wavelength is scanned
across the phase mask 36 to imprint the grating in the fibre 8. The
UV radiation is produced by an intra-cavity frequency-doubled argon
ion laser 42.
[0079] The phase mask 36 is positioned symmetrically between the
chucks 39, 40 so that the centre of the grating experiences zero
stretch. The piezo electric actuators are driven by an oscillator
43 at about 5 Hz by a triangular ramp signal ensuring that the
fibre 8 at the ends of the phase-mask 36 experiences a stretch
amounting to approximately half the period of the grating at those
points. In order to achieve a satisfactory symmetric half-period
stretch for a given length of grating, the following should be
satisfied: 3 f > v w
[0080] where f is the frequency of the stretching of the waveguide,
v is the scanning speed of the beam and w is the diameter of the
beam spot.
[0081] It is to be understood that this process can be repeated on
the same fibre at different, substantially contiguous locations
with the same phase mask to produce a long grating, in which case
apodisation by stretching will be applied asymmetrically at the
ends of the long pattern. To achieve this, the fibre may be
stretched by means of one of the piezoelectric devices only.
[0082] Alternatively, phase masks with different spatial
periodicities can be used to produce a chirped pattern. The
recorded patterns can be matched at their junctions by the
apodisation process.
[0083] FIGS. 17 and 18 show respectively the reflectivity and delay
of a chirp grating made according to the sixth example of the
present invention, in comparison with a theoretical prediction
produced by means of a computer simulation and the actual
performance of a corresponding unapodised grating.
[0084] Referring to FIG. 19, a WDM system which makes use of
apodised grating filters made in accordance with the invention will
now be described. A link in an optical communication system
comprises a WDM multiplexer 44 and an optical amplifier 45 at a
transmitter station. The output of the amplifier 44 is directed to
a 120 km length of optical waveguide 46. At a receiver station, an
amplifier 47 receives and amplifies optical signals from the
waveguide 46 and outputs them to a first port or an optical
circulator 48. A second port of the optical circulator 48 is
coupled to a bidirectional demultiplexer/multiplexer 49. A third
port of the optical circulator 48 is coupled to an optical receiver
50. The demultiplexer/multiplexer 15 is also coupled to four chirp
filters 51,52,53,54. Each of the chirp filters 51, . . . , 54 was
made by the method described with reference to FIG. 16 and are
adapted for compensating for dispersion of signals at respectively
1548 nm, 1552 nm, 1557 nm and 1562 nm. The dispersion parameter of
the chirp filters 51, . . . , 54 is 1600 p nm.sup.1. The operating
bandwidth of the chirp filters is .about.3 nm, allowing them to be
used over a temperature excursion of .+-.10.degree. C.
[0085] At the transmitter station, optical signals at 1548 nm, 1552
nm, 1557 nm and 1562 nm are combined and applied to the amplifier
45 by the multiplexer 44. The amplifier 45 amplifies the
multiplexed signals and launches them into the waveguide 46. During
their passage along the waveguide, the optical signals become
dispersed.
[0086] At the receiver station, the multiplexed signals are boosted
by the amplifier 47 and fed to the optical circulator 48 which
feeds them from its second port to the demultiplexer/multiplexer
49. The demultiplexer/multiplexer 49 distributes the component
signals of the multiplex signal to the chirp filters 51, . . . , 54
which reflect the applied signals in such a manner as to compensate
for the dispersion occurring in the waveguide 46. The compensated
optical signals are then recombined by the
demultiplexer/multiplexer 49 and fed back to the circulator 48
which outputs them at its third port. Finally, the compensated
optical signals are received by the optical receiver 50. With the
system of FIG. 19, using data rates in the region of 10 Gb
s.sup.-1, a total bit-rate x distance product of 4.8 Tb s.sup.-1 km
has been achieved. Also over 24 dB of cross-channel isolation has
been measured.
[0087] Many modifications and variations to the described examples
falling within the scope of the claimed invention are possible. For
example, gratings formed by the method of the invention may be used
in devices other than optical filters. The grating need not be
recorded in an optical fibre; it can be recorded in other forms of
optical waveguide such as a planar waveguide, or in a bulk optical
medium that is not necessarily configured as a waveguide. Also, the
grating need not necessarily be of a narrow elongate structure as
previously described. The waveguide elements could be disposed as
concentric circles, ellipses or other similar shapes, such that the
length of the grating extends radially outwardly of the recorded
pattern
* * * * *