OPTICAL SIGNAL SPLITTER
Field of the Invention
This invention relates to optical signal splitters for use in optical data transmission systems, for example optical communication systems and optical information processing systems.
Background Art
Compact, low-loss, dynamically reconfigurable optical signal splitters are critical to realising the potential of optical information processing systems and optical communication systems.
Optical communication systems such as fibre-to-the-home (FTTH) and fibre-to- the-business (FTTB) optical networks typically comprise a passive optical network (PON) architecture. A major bottleneck in conventional PON networks is the optical fibre intensive last section (or "last mile") of the network that connects to subscribers. One source of the bottleneck is that too many optical fibres must originate at the head end due to limitations in the number of times an optical signal can be split before it becomes too weak to use. Another source of the bottleneck in traditional PON networks having no active signal sources is that the maximum distance between the head end and a subscriber is usually limited to between 10 and 20 kilometres. In addition to these signal splitting and trunk length limitations, the adding/dropping of subscribers on a demand basis in conventional PON architectures typically requires physical rearrangement of data ports and optical fibres in the "last mile" of the network. Similar signal splitting bottlenecks also arise in conventional optical information processing systems such as optical computers.
Thus, a need exists for compact, low-loss optical signal splitters having active add/drop functionality.
Disclosure of the Invention
According to a first aspect of the invention, there is provided an optical splitter for use in an optical system, the optical splitter comprising:
one or more reconfigurable holographic diffraction gratings for selectively splitting one or more input signals into two or more split signals;
two or more optical pumps for generating two or more pump signals; and
two or more optical amplifier outputs for respectively receiving the split signals and the pump signals and amplifying the split signals to a predetermined power level in response to the pump signals;
wherein the reconfigurable holographic diffraction gratings are operable to selectively couple the split signals to the optical amplifier outputs.
Preferably, the optical splitter comprises a dynamically reconfigurable M:N splitter wherein M input signals are selectively coupled to N optical amplifier outputs by M reconfigurable holographic diffraction gratings.
Advantageously, the optical splitter is provided within a passive optical network. Preferably, the passive optical network is dynamically reconfigurable according to subscriber demand.
Preferably, the optical splitter further comprises a processor for controlling the one or more reconfigurable holographic diffraction gratings and a memory for storing programs to operate the processor in a prescribed manner to control the one or more reconfigurable holographic diffraction gratings.
Preferably, the optical splitter further comprises two or more detectors for respectively detecting the two or more split signals, wherein the detectors generate split signal control signals indicative of the power levels of the split signals, and wherein the processor controls the one or more reconfigurable holographic diffraction gratings in response to the split signal control signals.
Optionally, the optical splitter further comprises two or more detectors for respectively detecting the two or more pump signals, wherein the detectors generate pump signal control signals indicative of the power levels of the pump signals, and wherein the processor controls the one or more reconfigurable holographic diffraction gratings in response to the pump signal control signals.
Advantageously, the two or more detectors comprise a photodetector array.
Preferably, the two or more optical pumps comprise a vertical cavity surface- emitting laser (VCSEL) array. More preferably, the VCSEL array is a high power VCSEL array.
Preferably, the two or more optical amplifier outputs comprise a rare earth doped fibre array. More preferably, the doped fibre array comprises an erbium doped fibre (EDF) array.
Advantageously, the optical splitter further comprises one or more input optical elements to respectively couple the one or more input signals to the one or more reconfigurable holographic diffraction gratings. Preferably, the one or more input optical elements comprise one or more input collimating lens. More preferably, the one or more input collimating lens comprise an input microlens array.
Preferably, the optical splitter further comprises one or more relay optical elements for respectively routing and combining the two or more split signals and the two or more pump signals. More preferably, the one or more relay optical elements comprise one or more relay macrolens. Advantageously, the one or more relay macrolens are fixedly positioned in optical alignment by a rigid telescopic support.
Advantageously, the optical splitter further comprises two or more output optical elements to respectively couple the two or more split signals and the two or more pump signals to the two or more optical amplifier outputs. Preferably, the two or more output optical elements comprise two or more output collimating lens. More
preferably, the two or more output collimating lens comprise an output microlens array.
Advantageously, the one or more reconfigurable holographic diffraction gratings are generated by an optoelectronic very large scale integrated (VLSI) circuit chip. Preferably, the optoelectronic VLSI circuit chip includes the processor. More preferably, the optoelectronic VLSI circuit chip further includes the memory.
Advantageously, the VCSEL array comprises an optoelectronic integrated circuit chip. Preferably, the photodetector array comprises an optoelectronic integrated circuit chip. Advantageously, the VCSEL array and the photodetector array are integrated in an optoelectronic VLSI circuit chip.
Preferably, the optical splitter further comprises an optical substrate for respectively routing and combining the two or more split signals and the two or more pump signals.
In a preferred embodiment, the optical substrate comprises a planar surface and first and second spaced apart mirrors inclined at 45° relative to the surface and at 90° relative to each other, the pump signals and the split signals being arranged to propagate in spaced apart parallel opposite directions normal to the surface, the first mirror being configured to reflect the split signals to the second mirror in a direction parallel to the surface, and the second mirror being configured to transmit the reflected split signals in the surface-parallel direction and reflect the pump signals in the surface-parallel direction so that the reflected split signals and the reflected pump signals are co-aligned in the surface-parallel direction. Preferably, the second mirror comprises an antireflection coating for transmitting the reflected split signals and a reflection coating for reflecting the pump signals. More preferably, the photodetector array and the VCSEL array integrated in parallel spaced apart relationship in an optoelectronic VLSI circuit chip provided on the planar surface of the optical substrate, and wherein the first mirror is configured to tap a portion of the split signals off to the photodetector array. Advantageously, the optoelectronic VLSI circuit chip is attached to the optical substrate via flip chip bonding. Preferably, the output microlens array is
integrated in the optical substrate. More preferably, the output microlens array is etched in the optical substrate. Advantageously, the two or more split signals and two or more pump signals are respectively coupled to the doped fibre array by V- grooves integrated in the optical substrate.
Preferably, the optical substrate is glass or sapphire.
Preferably, the one or more reconfigurable holographic diffraction gratings comprise two-dimensional phase holograms.
According to a second aspect of the invention, there is provided an optical splitter for use in an optical system comprising:
one or more input microlens for collimating one or more input signals;
one or more two-dimensional reconfigurable holographic diffraction gratings for splitting the collimated input signals into two or more split signals, wherein the two-dimensional reconfigurable holographic diffraction gratings are inclined at an angle θ relative to the collimated input signals;
two or more reflective optical elements for reversing the direction of the split signals relative to the input signals by transverse internal reflection between the diffractive optical elements;
a two-dimensional VCSEL array and a two-dimensional photodetector array arranged in a spaced apart opposed relationship normal to the reversed split signals, the VCSEL array generating two or more pump signals;
a beam splitter arranged between the VCSEL array and the photodetector array, wherein the beam splitter couples pump signals from the VCSEL array to the reversed split signals and taps off a portion of the reversed split signals to the photodetector array so that the power of the reversed split signals can be monitored;
two or more output microlens for collimating the pump signals and the reversed split signals to a rare earth doped fibre array for amplifying the reversed split signals to a predetermined power level in response to the pump signals; and
an optical substrate, the input and output microlens and the beam splitter being integrated in the optical substrate, the reconfigurable holographic diffraction gratings being integrated in an optoelectronic VLSI circuit chip attached to the optical substrate, and the VCSEL array and the photodetector array being integrated in respective optoelectronic integrated circuit chips attached to the optical substrate;
wherein the reconfigurable holographic diffraction gratings are operable to selectively couple the reversed split signals to the rare earth doped fibre array.
Advantageously, the angle θ is 45° and the reversed split signals are parallel to the input signals.
Preferably, the one or more reflective optical elements comprise reflective diffractive optical elements.
Preferably, the optoelectronic VLSI circuit chip and the optoelectronic circuit chips are attached to the optical substrate via flip chip bonding.
Advantageously, the input and output microlens are etched in the optical substrate.
Advantageously, the two or more split signals and two or more pump signals are respectively coupled to the doped fibre array by V-grooves integrated in the optical substrate.
Preferably, the rare earth doped fibre array comprises an EDF array.
Advantageously, the optoelectronic VLSI circuit chip includes a processor for controlling the one or more reconfigurable holographic diffraction gratings.
According to a third aspect of the invention, there is provided a method of controlling the output power of split optical signals, the method comprising the steps of:
measuring the input power of one or more input signals;
splitting the input signals into two or more split signals using one or more reconfigurable holographic diffraction gratings;
measuring the split power of the split signals;
amplifying the split signals by a predetermined power gain;
determining a target output power level for the amplified split signals according to the measured input power and the predetermined power gain;
selectively controlling the output power of the amplified split signals at the target output power level by reconfiguring the one or more reconfigurable holographic diffraction gratings in response to the measured split power.
Brief Description of the Drawings
Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings in which:
Figure 1 is a schematic diagram of a generic optoelectronic optical signal splitter.
Figure 2 is a pattern for a 4-phase hologram and a corresponding replay field for a 1 :16 holographic beam splitter.
Figure 3 is a schematic diagram of a generic optical arrangement for optical signal collimation.
Figure 4 is a schematic perspective view of an optical splitter in accordance with a preferred embodiment of the invention.
Figure 5 is a schematic perspective view of an optical splitter in accordance with a preferred embodiment of the invention.
Figure 6 is a schematic perspective view of an optical splitter in accordance with a preferred embodiment of the invention.
Figure 7 is a schematic diagram of a generic fibre focuser array.
Figure 8 is a schematic perspective view of an optical splitter in accordance with a preferred embodiment of the invention.
Figure 9 is a schematic side view of an optical splitter in accordance with a preferred embodiment of the invention.
Best Mode(s) for Carrying Out the Invention
The general concept of optical splitting using available optoelectronic (or opto- VLSI) technology is shown in Figure 1. An input optical signal 10 is converted to a collimated optical beam 16 via a fibre collimator 12. The beam 16 illuminates an opto-VLSI chip 18, which acts as a phase-only diffraction grating, having a pattern which is reconfigurable. By loading an optimised multicasting hologram on the opto-VLSI chip 18, many beams of arbitrary intensities can be diffracted along arbitrary directions. These beams can be routed via a lens relay 22 to produce collimated optical output beams 26, which can be coupled into output optical fibre ports via a fibre focuser array.
The beam 16 illuminating the hologram is a collimated monochromatic beam having an intensity distribution which may be expressed as a Gaussian beam profile as given in the equation below: g(x,y) = -4Ge/(λ )
where AQ is the peak amplitude and / is a scaling factor which governs its total width.
For a Gaussian beam intensity, finite square hologram aperture, and finite circular Fourier lens, the replay field of a multicasting pixilated hologram is the superposition of sine functions propagating along different directions. This leads to minor sidelobe beams and hence crosstalk. To minimise the crosstalk, the hologram may be optimised, or the crosstalk signals may be routed such that they are not coupled into the output fibre ports.
An example of a multicasting hologram, h(x,y), is shown in Figure 2(a) which, in conjunction with a Fourier lens, produces a replay field, Ha(u,v), of 16 output beams at the focus of the Fourier lens, shown in Figure 2(b). For an 8-level phase hologram, it is possible to obtain an optical signal to crosstalk ratio greater than 20 dB.
Due to the utilisation of a Fourier lens after the hologram, the generated spots shown in Figure 2(b) are focused optical beams.
In order to obtain a low-loss or substantially lossless splitter architecture, it is necessary to amplify the signal beams. Therefore it is desirable that the collimated beams have a long Rayleigh length (distance along which the beam diameter increases by a factor of 2 ) such that they may be routed and combined with pump beams. Figure 3 illustrates how this is achieved utilising a lens relay 22 rather than a Fourier lens.
The lens relay 22 may be divided into two subsystems, namely: the macrolens relay 24, which transforms the diffracted beams that emerge from the hologram into focused beams, and the microlens array 25 which collimates the focused beams to the required beam diameter so that the focused beams 26 are efficiently coupled into a fibre focuser array.
A hybrid microlens-macrolens relay is preferable as it provides simple and efficient routing and collimating of the signal beams, and its flexibility enables adjustment of the stage-to-stage separation to the required values.
Figure 3 illustrates the propagation of three signal beams generated from an opto- VLSI chip hologram 18 at -4°, 0°, and 4°. The beams emerging from the macrolens subsystem 24 are parallel and focused. The microlens arrays 25 used after the macrolens relay 24 are placed at an appropriate distance in order to collimate the focused beams to the required diameter. This ensures that the beams enter the fibre focuser array with the correct numerical aperture and maximum allowable coupling.
A typical 1 :32 optical splitter has an insertion loss of approximately 15 dB. In order to compensate for such loss, optical amplifiers may be employed.
An optical amplifier is formed from erbium-doped fibre (of an appropriate length) which is spliced to each port of the output fibre collimator array. The fibre can compensate for the 15-dB inherent loss and the additional insertion loss of a 1x32 optical power splitter.
Figure 4 illustrates the architecture of an embodiment of a low-loss or substantially lossless optical splitter. The optical splitter 40 operates by splitting an input signal 10 into N collimated beams (N = 32 for standard passive optical networks (PON)). A VCSEL VLSI chip 18 integrates 980 nm VCSELs 28 for erbium-doped fibre amplifier (EDFA) pumping, 1550 nm photodetector array 30 for signal control, and glass, or any appropriate substrate, for optical beam routing and combining. The VCSEL pump beams, propagating within the glass substrate 32, are reflected by the right 45° tilted face and coupled into output EDF ports via a microlens array 27. In order to achieve signal beam coupling, 98% of the signal beams are first reflected off the left 45° tilted face, then combined with the pump beams using 1550 nm antireflection coating (to minimise surface reflection losses) on the external surface of the right 45° tilted face, then the combined red and blue beams are coupled into the output ports via the microlens array 27. Signal monitoring can also be achieved by integrating a 1550 nm detector array 30 that
picks-off small percentages (for example, 2%) of the signal beams so as to control the output optical powers as shown in Figure 4.
By controlling the holographic pattern on the opto-VLSI chip, the input beam can be arbitrarily split and coupled into any output EDF port, which compensates for the splitting loss, leading to low-loss or substantially lossless optical splitting. For an input EDF signal less than -10 dBm, an optical gain of more than 15 dB can be achieved with typical VCSEL pump powers.
Figure 5 shows an embodiment of an optical splitter in which the input fibre and collimation lens are integrated in a single steel barrel. Fibre collimators featuring very low insertion and return losses are commercially available. The collimating lens is a macrolens or a micolens depending on the required collimated beam diameter. Alternatively, the collimating lens is an 8-16 level diffractive element on glass, so as to minimize aberration. In order to ensure a robust relay, wherein all elements will remain stable and well aligned and removal, replacement and repositioning is allowable, all macrolens are advantageously aligned, appropriately spaced, and centered at a 90-degree angle to the optical axis. Further, the microlens are advantageously mounted at the right angle and with a high degree of position accuracy. Figure 5 illustrates a possible implementation of the lens relay 22, whereby a telescope-like assembly is used to ensure the lens alignment and positions are highly accurate.
The particular alignment of the preferred embodiment of the microlens array 25 and the VCSEL chip 29 is illustrated in Figure 6. The collimated signal beams 16 are first reflected from the lower 45° tilted surface by using 1550 nm reflection coating, then combined with the VCSEL pump at the upper 45°-tilted surface which is externally antireflection-coated for 1550 nm. By integrating a microlens array 27 directly into the glass substrate 32 it is possible to couple the pump and signal beams into the EDF ports 34.
The coupling of the pump beams and signal beams is performed by using a fibre focuser array which has been integrated onto the glass substrate via V-groove technology. A general structure of a focuser array is shown in Figure 7. A thick
epoxy may be deposited between the microlens array 27 and the fibre array 34 so as to achieve refractive index matching.
Typical characteristics of available fibre focuser arrays are as follows:
Lenslet type: Epoxy on glass or fused silica Lens diameter/pitch: 0.1-2.0 mm
Pitch accuracy: 0.5 micron Working distance: 10-500 mm Insertion loss: 0.5 dB Return loss: -50 dB Operating temperature: -40-85 °C
A further advantage is in the flexibility of the opto-VLSI chip to partition to sub- holograms. In preferred embodiments, such partitioning allows the realisation of either a multi-port low-loss or substantially lossless optical splitter, or a many fold increase in the number of output ports. For example, dividing the active area of an opto-VLSI chip into 4 pixel blocks, enables the use of each pixel block as a beam splitter, resulting in a 4-fold increase in splitting capacity.
To achieve such splitting, it is necessary to modify the architecture slightly. Firstly, the input fibre collimator is replaced by a collimator array, which consists of integrated discrete elements.
Secondly the opto-VLSI chip is divided into pixel blocks, each pixel block splitting the input beam into a linear collimated beam spot. The same macrolens relay is used as previously, however as shown in Figure 8, the relay's microlens array becomes a two-dimensional array 45 and the tilted surfaces of the VCSEL glass substrate 32 are elongated in the lateral direction so as to accommodate a greater number of beams.
An alternative embodiment, shown in Figure 9, utilises the same concept used in Figure 4, except that a transparent optical substrate 101 replaces free space for beam propagation. This elimination of free space in the design has the advantage of allowing the device to be more robust. An input light beam 102 enters the device through the optical fibre 103 and is collimated by a first lens 104 that is etched directly into the optical substrate, and then the collimated beam 105 is split
by the flip-chip bonded opto-VLSI processor 106 in the same manner as in the embodiment shown in Figure 4. The normal axis of the opto-VLSI processor 106 is tilted an angle θ with respect to the direction of the collimated beam 102. The optimum tilting angle θ that leads to a minimum substrate size is 45°. Unlike the previous embodiment, however, the diffracted beams 107 are now guided inside the substrate 101 by total internal reflections from the reflective diffractive optical elements (DOEs) 108, 109 and 110 mounted on the surface of the optical substrate 101. The DOEs 108, 109, and 110 are equivalent to the lens relay shown in Fig. 3. An optimum design of these DOEs allows separation of the diffracted beams a maximum amount, depending on the maximum steering angle of the opto-VLSI processor 106. After the internal reflections, the beams strike the reflective DOE 111 mounted on the surface of the substrate 101. This DOE 111 is designed so that the diffracted beams 107 are collimated up at a right angle along the optical axes of the output collimators 118. 98% of the power of these collimated beams are coupled into the output erbium doped fibres EDF array 119 via the collimator array 118, whereas a monitored signal beam 113 (2% of the beam power) is routed via the coated optical splitter 112 to the photodetector array 114, that is flip chip bonded on the surface of the substrate 101. The role of the photodetector array 114 is to monitor the signal powers coupled into the input ports of the EDF array 119. The two-dimensional VCSEL array 115 is also flip- chip-bonded to an ultra-thin semiconductor (UTS) chip 116, which is a semi- transparent photodetector array attached to the surface of the substrate 101 to sense small fractions of the powers of the pump beams 117 generated by the VCSEL array 115, for monitoring the optical gains of the ports of the EDF array 119. The VCSEL pump beams 117 generated by the two-dimensional VCSEL array 115 are reflected up at a right angle by the coated optical splitter 112 and coupled to the EDF ports to provide optical amplification for the signal beams.
It will be apparent from the above description that preferred embodiments of the invention provide an intelligent lossless optical splitter that employs the multicasting capability of opto-VLSI chips and uses VCSEL pump arrays in conjunction with erbium-doped fibres to compensate for the splitting loss. Preferred embodiments facilitate dynamic access control of FTTH subscribers,
hence allowing a much improved service management to be attained, in comparison to the service management of current passive optical networks. By optimising the holographic diffraction grating on the opto-VLSI chip, an incident beam can be split to many output beams of arbitrary intensities. This feature of the opto-VLSI technology is used to provide dynamic subscriber access control, and hence flexible service management in FTTH optical networks. In addition, by partitioning the opto-VLSI chip into several pixel-blocks (e.g. 4) in preferred embodiments, each hologram on a pixel block is optimised to work as an independent 1 :N intelligent optical splitter, allowing dense and cost-effective passive/active optical network to be realised.
It will be apparent that the advantages of preferred embodiments of the intelligent lossless optical splitter include:
(1) optical splitting with passive and active options, for future FTTH networks, (2) amplification, which improves quality of service and increases last-mile spans,
(3) formulated as a static device, therefore, it is reliable, and
(4) cost-effective since all ports share a single opto-VLSI chip and a single VCSEL pump chip. With mass production a substantial reduction in splitter cost is achieved by preferred embodiments.
The embodiments have been described by way of example only and modifications are possible within the scope of the invention.
Throughout the specification, unless the context requires otherwise, the word "comprise" and variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers.