US 20070230854 A1
An integrated active electrical waveguide is described for optical waveguide modulators. The active waveguide may include an optical waveguide, a plurality of modulators integrated into the waveguide to modulate the index of refraction of the waveguide, and a plurality of local amplifiers, each coupled to a modulator.
1. An active waveguide comprising:
an optical waveguide;
a plurality of modulators integrated into the waveguide to modulate the index of refraction of the waveguide, each modulator having a source on one side of the waveguide and a drain on the other side of the waveguide;
a voltage driver; and
a plurality of level shifters each coupled to the voltage driver and to a modulator to drive the voltage from the voltage driver to one of the plurality of modulators.
2. The active waveguide of
3. The active waveguide of
4. The active waveguide of
5. The active waveguide of
6. The active waveguide of
7. The active waveguide of
8. The active waveguide of
9. The active waveguide of
a second optical waveguide;
a second plurality of modulators integrated into the second waveguide to drive an electrical field into the waveguide, the electrical field having a polarity opposite that of the first electrical field; and
a plurality of local amplifiers, each coupled to a modulator.
10. The active waveguide of
11. The active waveguide of
12. An optical transmitter comprising:
an input data retiming circuit;
an optical source;
an optical waveguide to receive light from the optical source and couple it into an output optical fiber;
a plurality of phase modulators integrated into the waveguide to drive an electrical field into the waveguide to modulate the light from the optical source, each modulator having a source on one side of the waveguide and a drain on the other side of the waveguide;
a voltage driver; and
a plurality of level shifters each coupled to the voltage driver and to a modulator to drive the voltage from the voltage driver to on of the plurality of phase modulators.
13. The transmitter of
14. The transmitter of
15. The transmitter of
16. A method comprising:
forming a polysilicon optical waveguide on a silicon substrate;
forming a plurality of oxide modulators on the silicon substrate to modulate the index of refraction of the waveguide each modulator having an oxide source on one side of the waveguide and an oxide drain on the other side of the waveguide;
forming a voltage driver on the silicon substrate; and
forming a plurality of metal oxide semiconductor level shifters on the silicon substrate, each coupled to the voltage driver and to a modulator to drive a respective modulator.
17. The method of
18. The method of
The present description relates to optical waveguides for optical communications and, in particular, to driving a phase modulator to compensate for the inherent capacitance of a waveguide.
Electro-optic modulators find significant use in fiber optic communication systems to modulate data onto a light beam. Light is conducted through a waveguide and electrical signals are applied across the waveguide to apply a modulation to the light beam. Lithium niobate is a common material for such modulators because it works well as a waveguide for light and also exhibits excellent properties for modulating light with electrical inputs. However, lithium niobate is not a good material for semiconductor circuitry. As a result, a lithium niobate modulator requires a separate chip, typically silicon, to carry circuitry for the phase modulator. Electrical wires connect the two silicon chip outputs to the lithium niobate chip. In addition, modulating light with a lithium niobate modulator requires typically a swing of 5V; a swing that modern CMOS processes cannot easily supply. A much less expensive modulator may be made by putting both the circuitry and the waveguide on the same chip, however, silicon is not a good material for a light waveguide.
In order to have a long interaction length between the electrical signal and optical wave in an electro-optic modulator, most electro-optic modulators include a long waveguide. An example of such a waveguide is a Mach-Zehnder lithium niobate modulator. Such a modulator will normally have electro-optic material (i.e. lithium niobate) embedded in an electrical waveguide with the modulating electrical field penetrating the electro-optic material of the optical waveguide. The electrical waveguide will typically have a characteristic impedance of about 50 Ohms and can be driven by a 50 Ohm driver.
Semiconductor modulators constructed similar to a Mach Zehner lithium niobate modulator have a much higher capacitance per unit length. This creates a much higher electrical load and a much slower phase velocity. The low phase velocity can cause a phase mismatch between the optical and the electrical wave as the two waves travel along the waveguide. One effect of the phase mismatch is to limit the maximum data rate of the communication system that uses the semiconductor modulator.
It is very difficult to change the inductance per unit length of a semiconductor waveguide by changing its geometry, that is the shape, depth, width, or length of the waveguide. The inductance is only weakly dependent on the geometry of the waveguide and normally will remain within a range of 1 to 1.3 nH/mm. Changing the waveguide geometry also has other significant effects on the modulator and the geometry cannot be changed freely without considering the many other factors that are affected.
Embodiments of the present invention will be understood more fully from the detailed description given below and from the accompanying drawings of various embodiments of the invention. The drawings, however, should not be taken to be limiting, but are for explanation and understanding only.
Active components, that is electrical driver stages may be made an integral part of an electrical waveguide. The optical components and the active electrical components may be integrated in one process on a single semiconductor substrate or placed in close proximity. A waveguide modulator and all of its drive circuitry may be formed on a single silicon substrate using a bipolar junction CMOS (Complementary Metal Oxide Semiconductor) process.
As mentioned above, such a silicon waveguide has a high capacitance C per unit length. The high capacitance makes the waveguide a high load Zo. The load may be determined by Zo=(L/C)1/2. The high capacitance also creates a high phase velocity vph through the waveguide. The phase velocity may be determined by vph=2π/(LC)1/2.
In one embodiment, and as shown in
The charge carriers on the polysilicon gate change the refractive index of the waveguide. The source and drain of the transistors may be grounded. The change in the refractive index may be used to shift or modulate the phase of the light in the waveguide. Such phase shifters may be used to modulate data onto the light wave and also to compensate for the high capacitance. While the present invention is described in the context of CMOS process and structure, embodiments of the invention may be adapted to other technologies and materials.
An active driver circuit can be implemented in a variety of different ways to reduce the capacitance of the waveguide and reduce the phase velocity of light traveling through the waveguide. One simple way is to put a level-shifter (not shown) implemented as a source follower or emitter follower between the driver 16 and the transistors 22. This decouples the load capacitance from the waveguide. The level-shifter works as an impedance transformer. This increases the phase velocity and reduces the phase mismatch between the optical wave and the electrical wave as light travels across the waveguide.
In such an implementation, there may be power lost in the termination resistor in proportion to the square of the voltage provided by the driver. In addition, the input capacitances of the level-shifter may be non-linear across the voltage swing. This may lead to distortion in the optical signal conducted through the waveguide. An impedance transformer stage may be implemented in a compact way with a few components or in a complex way with many additional components and features (not shown). In one embodiment, the level shifter uses at least two transistors.
As compared to the relatively high capacitance per unit length of the waveguide, the drivers 40-1 to 40-n may be designed to have a much lower input capacitance. This leads to a higher characteristic impedance for the circuit as a whole and, as a result, a higher resistance termination resistor. The higher resistance results in less power being wasted than in for example, the
Considering the waveguide and coupled transistors in more detail, the device is based on the free-carrier plasma dispersion effect, which may be exploited to create an electrically controllable optical path length for photons with energies below the bandgap energy. In silicon, the bandgap energy is approximately 1.1 eV, and the waveguide is operated with photon energies of roughly 0.8 eV. In addition to altering the effective path length of the waveguide, the free-carriers absorb some of the light in the waveguide. This effect is manifested as a small voltage-dependent loss (VDL).
The Table provides test results that demonstrate how the propagation speed of an optical signal is affected by the driver for the transistors coupled to a silicon waveguide. The values of the Table are taken from a design with 0.25 μm CMOS and 70 GHz fT bipolar BiCMOS process applied to a silicon-based Mach-Zehnder modulator.
Several different waveguide versions were tested for the data of the Table. The passive waveguide, version A, corresponds to a waveguide with no driver circuits or coupled transistors as shown, for example in
The characteristics of the different implementations of the electrical waveguides include inductance per mm of waveguide (L/mm) as suggested by the inductors L1, L2, L3, to Ln in the figures, capacitance per mm of waveguide (C/mm). Vph, the phase velocity of the electrical wave in meters per second, Z0, the characteristic impedance of the waveguide, and Δτ/4 mm, the propagation time mismatch between the optical wave (vopt=8.5*107 m/s) and the electrical wave for the tested (4 mm long) waveguide.
Referring to the Table, for the tested 2.5 Gbps modulated signal with a period of 400 ps the Δτ through the passive waveguide (version A) represents almost a full period. As a result, at the end of the waveguide, the electrical signal does not contribute constructively to the optical signal at all. A rough estimation of the maximum acceptable Δτ is about 200 ps for 2.5 Gbps and 50 ps for 10 Gbps. Based on this estimate, the waveguide with level-shifter (versions B and C) is suitable for data rates up to 2.5 Gbps and the waveguide with integrated amplifier (version D) for 10 Gbps or higher.
As for the other features of
The system of
The physical interface between the driver and modulator gives a low inductance while still allowing for a simple wire-bonding technique. Each section of the waveguide has a bond pad for the two differential signals (applied to the p-type poly-Si gates of the phase shifters) and a bond pad for the RF return path (connected to the modulator n-type substrate).
The optical interferometer bias point may be controlled by voltages taken directly from the PCB that carries the driver chip and modulator chip. This includes the modulator substrate direct current (dc) voltage, which may be connected to the lowest potential in the system VEE. For the phase shifters of
A top undoped polysilicon layer is formed over the tops of the oxide ridges and metal contact pads 80, for example aluminum pads, are formed on either side of the waveguide channel over the polysilicon layer. The two oxide ridges form a source and a drain for a transistor across the gate waveguide. The voltage between the two oxide layers generates an electric field across the waveguide. By placing the metal contacts to the side rather than on top of the rib waveguide 78, the optical absorption due to the metal contacts is almost eliminated.
The oxide regions 76 on either side of the waveguide maintain optical confinement and prevent the optical field from penetrating into the areas where the metal contacts are located. To make Ohmic contacts to the metal, both the crystalline silicon slab and poly-silicon may have a surface doping concentration of 1×1019 cm-3 The particular structure of the waveguide and integrated transistor may be modified to suit different applications and to accommodate advancements and improvements.
The active drivers may further be used to adjust for different modulators, or modulators made from different semiconductor materials, for example any compound semiconductors, such as InP or GaAs, among others. The active drivers may be set with values to compensate waveguides with a wide range of different capacitances or impedances and to accommodate modulators that have an optical waveguide formed of other materials to suit a variety of different applications.
The parallel branch 95 is coupled to the negative output 99 of the driver which connects to a sequence of drivers 107-1 to 107-n to send a voltage to the coupled transistors 103-1 to 103-n as described above in the context of
A lesser or more complicated waveguide, driver circuits, modulator or transmitter may be used than those shown and described herein. Therefore, the configurations may vary from implementation to implementation depending upon numerous factors, such as price constraints, performance requirements, technological improvements, or other circumstances. Embodiments of the invention may also be applied to other types of systems that use different inputs and outputs than those shown and described herein.
Many of the methods and apparatus are described in their most basic form but steps may be added to or deleted from any of the methods and components may be added or subtracted from any of the described apparatus without departing from the basic scope of the present invention. It will be apparent to those skilled in the art that many further modifications and adaptations may be made. The particular embodiments are not provided to limit the invention but to illustrate it. The scope of the present invention is not to be determined by the specific examples provided above but only by the claims below.