Thermal auto-tuning and stabilization of optical ring resonators

Abstract

An optical micro-ring modulator (MRM) requires temperature tracking and locking such that the resonance notches / peaks of the MRM device are actively thermally shifted to coincide with the narrow input optical wavelength. Accordingly, it would be beneficial to auto tune an MRM device so that it functions as a high-speed modulator and accommodates initial process variations in manufacturing, the random thermal heating and cooling of the environment, and drift of the input laser line wavelength. The control system generates an error signal by comparing complementary tapped optical signals from the through and drop output waveguides of the MRM device where the controller is tuned to the high-speed electrical modulation of the device along with the slow-speed thermal transient behaviour of the MRM device. The controller can operate using balanced electrical modulation code, non-balanced modulation code and tuned to optimize for optical amplitude modulation (OMA) and extinction ratio.

Description

THERMAL AUTO-TUNING AND STABILIZATION OF OPTICAL RING RESONATORSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of priority to U.S. Provisional Patent Application 63 / 550,369 filed February 6, 2024; the entire contents of which are incorporated herein by reference.TECHNICAL FIELD

[0002] The present disclosure relates to devices that perform optical signaling, particularly in the data communications and sensing fields, and the control of these structures using simple electronic circuits to regulate performance as it relates to inherent physical process variation of devices during manufacturing, random operational changes in internal and external temperature along with possible changes or drift in the incident wavelength of light that they use over time.BACKGROUND

[0003] The background for this invention is in the field of optical devices and their use in both passive optical filtering for multiple wavelength systems and in data modulation where electrical signals are converted into optical signals and back again. There have been a vast number of devices and structures developed over the past decades that are related to both the sourcing of light (such as the LED and Laser) but also to the modulation of that light by either direct or indirect means. Directly modulated light typically means that the same flow of current that generates the light (the photons through some form of excitation and spontaneous emission) is also modulated in a varying pattern to change the intensity of the light (although phase modulation has also been shown). Whereas indirect modulation means that the light source is kept constant and a means of interrupting the light to create pulses or phase delays is done with a second device from the light source.

[0004] There are numerous devices that use direct modulation, such as the slower light emitting diode (LED) as well as the Fabry-Perot (FP) laser, the vertical cavity surface emitting laser (VCSEL) diode, and the distributed feedback (DFB) laser. The modulation rates of these devices are increased every year while confronting obstacles from the electrical properties ofthe materials and interconnects. These speed increases become ever more challenging to obtain because of ever higher current densities in smaller areas, higher thermal variations and high- frequency deterministic and stochastic output emission. The main reason for this is the interaction between the moving charges inside the ever-smaller devices to produce the light modulation. To increase modulation rates, physical sizes tend to have to decrease which in turn decrease the so-called electrical capacitance, resistance and sometimes inductance that comprise the frequency dependent characteristics of attenuation or resonating circuits that make up the device. In addition are the thermal effects in the devices that alter carrier densities that can cause unwanted optical performance to occur.

[0005] Although directly modulated devices will undoubtedly remain relevant in the world for the foreseeable future, the set of device structures that use indirect modulation to encode electrical data (or pressure or temperature signals for sensor devices) into optical data appears to be set to advance the state of the art in data transmission well into the future as well.

[0006] The indirect modulation technique can involve structures that modulate the output amplitude, such as simple binary non-return to zero (NRZ) a version of on-off keying (OOK). Indirect methods are also used for more complicated encoding involving multilevel amplitude signaling such as pulse amplitude modulation (PAM-) 4, 8, 16 etc. levels. This method encodes for more than 1 bit in a single symbol, for example in PAM-4 each symbol can represent 2 bits as each symbol can represent 1 of 4 levels bit-pairs (i.e.: 00, 01, 10, or 11). This method is largely implemented when the frequency bandwidth of the channel cannot be increased, but the multiple amplitudes can have a large enough range between them to be discerned in the presence of noise.

[0007] More complicated methods of optical indirect modulation borrow from the radiofrequency (RF) domain by using the long optical coherence of laser signals and using phase delay techniques of carrier frequencies to encode relative changes in the phase of a signal relative to an unmodulated carrier. Examples of these techniques are encoding schemes like phase-shift keying (PSK) or quadrature amplitude modulation (QAM) which involve multiple slices of phase-differences between two coherent signals. These methods are often much more immune to noise than the amplitude methods, although their drawback is that the nature of the receiver elements is usually much more complex and tends to use several photodetector elements for each signal with complicated and power hungry digital signal processing (DSP) to extract data signals from potentially very corrupted received signals - especially when used for very long-distance transmissions in optical fiber.

[0008] Over the past decades many of these devices have been reduced in size and power as device processing capabilities have been refined. One of the earlier devices was an electroabsorption modulator (EA modulator) and was a way to absorb a light beam inside a material like a Gallium Arsenide quantum well structure. Electro-optic materials on the other hand use electrical fields and carrier densities to change propagation lengths of an optical signal by changing the density of states in a material. A particularly well-known integrated device for this effect is the Mach-Zehnder optical waveguide interferometer. Although originally the MZ interferometer was the size of a table and was used to measure very minute changes in distance by using free-space optical beams of light along two rectangular paths that used optical beam splitters and combiners to create an interference pattern on a flat screen. The planar waveguide MZI essentially does the same thing only using optical guides on a planar light chip (PLC) or photonic integrated circuit (PLC) and has been miniaturized to the size of a microchip.

[0009] The indirect modulation method using on-chip optical structures such as the MZI is normally based on a form of optical interference. To create optical amplitude modulation, the physical overlap in a region of two optical signals with relative phase delay that either produce constructive interference; resulting in a high amplitude output signal or destructive interference; resulting in a low amplitude output signal. The MZI structures normally have 2 paths, one path typically is subjected to an electrical field. The material used can be a semiconductor or crystals with dual axis internal electric fields such as Lithium Niobate (LiNOs). Lower voltages of 1 or 2 V can be used on the longer the path, where the induced phase change in one of the arms can produce constructive / destructive interference at the output of the device. However, these devices can be quite large, several millimeters in length, and their integration and miniaturization is limited. However, with careful design of travelling wave electrical transmission lines with 50 Ohm terminations, these devices can operate at speeds well above 200-Gbps. Furthermore, their sensitivity to temperature and physical fabrication variations is small, and they can be demonstrated to work for many commercial applications.

[0010] While these physically larger devices are being deployed and satisfying data transmission requirements today, there are other devices, such as active optical Micro Ring Modulator (MRM) technologies that have unique indirect modulation characteristics but may prove to be far more scalable than the MZI technology. First, the MRM device is much smaller, in the order of a few microns, and thus many more could be integrated into a single chip. Secondly, the structures also easily provide optical filtering techniques as well such as wavelength selection for wavelength division multiplexed (WDM) laser lines.

[0011] The Optical Micro Ring Modulator (MRM) device structure is a straight optical guide path with a secondary circular optical guide path placed beside, but not touching, the first straight optical guide path. Through the process of weakly guided mode coupling, light in the straight path can build in the circular path that subsequently is allowed to interfere with itself as it propagates back around the circle. The MRM uses constructive and destructive interference, although it does so more by tuning to a particular physical geometry related to the wavelength of light to create a resonant peak and / or notch for the intensity of the output light at a specific wavelength. In other words, depending on the diameter of the circular guide and the coupling ratio, a specific wavelength (broadly speaking) of light will have exactly the right propagation length to interfere with itself around the circular path to pass (or be dropped) with all other wavelengths in the path to remain unaffected. Of course, in real conditions, the filter is not infinitely sharp and has a non-zero bandpass opening. Also, because the quotient of path length and wavelength can have multiple integer values, the MRM also has the property that there are multiple pass / notch frequencies repeated for multiple wavelengths. This gives rise to a repeating pass / notch pattern over wavelength call the Free-Spectral Range (FSR). This is unlike the MZI that uses 2 longer parallel paths of delayed light to create an output interference state at the output.

[0012] It is also useful to note that the MRM structures can be more complicated than the single bus and ring structure. Different filtering and modulating structures can use more than 1 ring, coupled to other rings. They can have bus structures that conform to parts of the ring, and the ring itself may have straight sections like a race-track oval, among other shapes. The materials that comprise the MRM structures as well can be made of more exotic materials such as chalcogenides or optical amplifier elements such as Erbium and detector elements such as Germanium for certain wavelengths. Even the vertical structures can involve resistive materials as well as air-pocket undercuts to allow local heating properties to be improved.

[0013] The MRM can be a passive device, and made of non-semiconductor material such as glass (SiO2) or silicon nitride (SiN), and its characteristic peak / notch will be dictated by the physical geometry and the temperature of the device. This material system typically has better propagation with lower loss per millimeter. Typically, these passive devices can also be tuned for their resonant pass / notch wavelength by using resistive micro heaters patterned above or around the ring portion. A small input electrical current can heat the device and thus thermally expand minutely the structure changing the specific wavelength of the pass / notch.

[0014] However, the MRM can also be made using a semiconducting material like silicon (Si). It is well known that for Near Infrared wavelengths, silicon is fairly optically transparent.Although it does have an optical loss or absorption coefficient of a few dB per centimeter, optical guides patterned out of silicon can act similar to glass guides. The most important aspect of this is that the silicon can also be doped to make active devices such as PN-junctions that can be patterned inside the optical guides themselves. Hybrid systems (such as Si and SiN) which combine different materials to take advantage of different beneficial properties have also been described.

[0015] The nature of the optical wavelength interference effect that the MRM device uses, causes it to be sensitive to temperature and other slight changes in physical geometries as well as voltages when active devices like PN junctions are used. In some cases, this can be advantageous (as in sensing devices), but can lead to difficulties controlling and maintaining specific operating points. It is possible that an MRM device, used as a modulator, can be rendered inactive just because the input laser wavelength and the notch / peak of the MRM device do not align with each other. This can happen as temperature changes and the optical waveguide material heats or cools causing the notch / peak to move in wavelength by only nanometers. In a standard 10-um diameter MRM device made in silicon, the thermal drift can be as much as 0.07-nm / °C. Whereas a reverse bias voltage signal applied to the PN junction within the MRM structure may only be able to move the notch / peak 0.010-nm with a 1-V swing. Of course, temperature dependence can also alter or distort performance for sensing devices if these changes cannot be monitored and / or corrected as well.

[0016] Processing variations where side walls of different structures that are etched too little or too much on the same chip result in a collection of devices that may in fact be “out of tune” with each other relative to the initial design requirements will also cause poor or no performance if not compensated-for.

[0017] Regardless of the mechanism for drift in the MRM device, many prior-art solutions have attempted to analyze an optical output signal from an MRM device and produce a counterbalancing thermal heating (with passive cooling) regulator near the MRM device to attempt to lock in an optimal set-point. These systems have shown a great deal of promise and success overall and have been able to achieve a lock-in for these modulators that maintain optimal optical extinction ratio (ER) and / or optical modulation amplitude (OMA).

[0018] However, as these systems have only typically focused on stabilizing a single MRM device at a time, there has been no limit to the amount of supporting electronics used to implement the closed-loop control system. The closed loop control has been able to use the full suite of analog -to-digital (ADC) conversion chips, as well as the full processing capability of microcontrollers or field-programmable-gate-arrays (FPGAs) to implement digital signalprocessing on digitized signals. The output of the control electronics is connected to simple resistive heaters buried in the substrate around the MRM device. As mentioned earlier, the heating of the MRM device changes the material properties of the silicon waveguide, allowing the index of reflection in the semiconductor to change and resulting in a shift in position of the notch.

[0019] However, in each of these cases, the focus has always been to produce the largest and most pronounced optical modulation amplitude (i.e.: the highest optical power logical “one” and the lowest optical power logical “zero”). This normally requires that the lowest point in the notch be found so that any modulation about this point should result in the highest change in optical power. The control system used for the comparison, assuming this local minima of the notch (or maxima peak), requires the set point be compared to a known reference value that is in fact unknown at the start of the control process. These algorithms therefore require a user- defined “set-point” that must first be “found” and then “stored” in a memory to compare against after finding the optimal performance - all assuming this optical set point does not change over time.OBJECTIVE OF THE INVENTION:

[0020] In this invention, it is described a method that uses a differential measurement technique for an MRM device that automatically achieves a lock-in point without the need for an absolute reference nor a memorized state to compare against. This lock-in state can be designed to be very close to the previously mentioned “optimal” state of the local minima of the notch (or maxima of the peak), but it is far easier to obtain, while still providing for an optical output with a large extinction ratio and optical modulation amplitude for a given input optical power.

[0021] Furthermore, a second objective of the invention is to be able to make simple enough control circuits, with the fewest number of elements, so that many tens, hundreds, or thousands (or more) MRM devices can be actively tuned at the same time on the same chip.

[0022] In its simplest form, the control circuit is based on a simple servo “PID” controller (proportional-integral-derivative control) that uses simple circuits and filtering to achieve a fully analog feedback controller that does not require a defined set point.

[0023] The importance of a simple control circuit is highlighted by future requirements for data modulation and transmission that will require many parallel MRM devices simultaneously. If many MRM devices must be used, the overhead of existing control circuit architectures each requiring ADC and DSP using microcontrollers or FPGAs to control single MRM devicesbecomes excessively large and power hungry. In applications such as parallel channel transceivers, in photonic matrix multiplication using multiple layers of MZ interferometers, and in applications such as parallel optical logical structures where many tens or even hundreds of devices might be required simultaneously, simple control systems acting on each device will be an absolute necessity.

[0024] The inventions described herein are therefore primarily related to designs and methods of stabilize MRM devices while being exposed to internal and / or external temperature changes. However other sensitive optical devices, such as the MZI device previously mentioned may also benefit from these improvements. In addition, these same methods can also help mitigate fabrication process variations, such as waveguide etch depths, or minute lithographic errors in producing the devices, but may also aid in cases where the incident optical wavelength drifts over time.

[0025] Other aspects and features of the present invention will become apparent to those ordinarily skilled in the art upon review of the following description of specific embodiments of the invention in conjunction with the accompanying figures.EMBODIMENTS

[0026] According to an embodiment, there is provided an optical device called a micro ring modulator (MRM) or resonator (MRR) that is composed of straight and circular waveguides on a planar light chip (PLC) or photonic integrated circuit (PIC). This class of device can modulate light intensity by using an interference effect and a modulated change in electron / hole carrier density to effect a change in optical propagation path to function as an optical amplitude modulator. However, this device requires a thermal control system to lock-to or track the input narrow linewidth optical wavelength source of light to the device’s resonance notch / peak for proper operation.

[0027] In one embodiment, the optical device thermal control system will measure a portion of the complementary tapped optical outputs from the through and drop ports of the device producing two complementary photocurrents.

[0028] The optical device thermal control system will actuate a thermal resistance nearby to the device to increase the localized temperature.

[0029] The increase in localized temperature will shift the resonance notches / peaks of the device towards lower wavelengths, so called red-shifting.

[0030] The red-shifted notches / peaks of the device will then align with the narrow linewidth input optical wavelength source of light by the action of negative feedback from a proportional- derivative-control (PID) controller that uses the complementary optical signals to produce a subtractive error signal without the requirement of any external reference signals nor calibration nor memorized state.

[0031] As an electrical to optical modulator device, an input electrical modulation signal will be used to modulate the notches / peaks of the modulator. This modulation signal maybe be balanced code such that there are on average as many 0’s as there are l’s to ensure that lower- speed average tracking of the PID can be accomplished.

[0032] In a second embodiment, the PID controller is modified to accept a non-balanced set of input electrical stimulus thereby allowing completely random bit patterns to be used including mostly 0’s or mostly l’s.

[0033] In a second embodiment, the PID controller is upgraded to allow for a variable yet complementary gain for each of the optical tapped signals.

[0034] This variable complementary gain allows a weighted shift of the error signal depending on the ratio of 0’s to l’s in the signal.

[0035] The variable complementary gain can also be tailored matched to the characteristics of the device such that a linear function or other higher order functions (such as quadratics) can be used as the gain function.

[0036] In a third embodiment, the gain for each of the optical tapped signals can be fixed, yet calibrated to be asymmetric to allow for a static adjustment to the optimize the optical amplitude modulation (OMA) and the extinction ratio (Er) of the modulation.

[0037] In a fourth embodiment, and only as a limitation to the availability of integrated devices for cooling a device on chip, the controller uses active heating but passive cooling. Although with any future advent of device level active cooling, all embodiments can be augmented with active heating and cooling.

[0038] In a fifth and sixth embodiment, to address the limitations of active heating with passive cooling, a device may not be able to lock to a point if the notches / peaks need to be cooled at first, or if the external temperature of the system becomes hotter than where the initial locking point is found, a so-called “dead-zone”.

[0039] The PID controller can be further adapted to create a small voltage increase if the error signal ever becomes less than zero which forces the notches / peak to be red-shifted momentarily. This forces the PID controller to then attempt to lock onto the next set of notches / peaks of the device’s free-spectral range (FSR).

[0040] The PID controller can also be adapted to include a form of absolute value rectification when the error signal becomes lower than zero, if the PID control is made with symmetric power supplies above and below zero, the locking point of the PID controller can be the alternative point to the left of the notches / peaks with the caveat that all bit patterns are inverted.

[0041] Other aspects and features of the present invention will become apparent to those ordinarily skilled in the art upon review of the following description of specific embodiments of the invention in conjunction with the accompanying figures.DESCRIPTION OF THE DRAWINGS

[0042] Further features and advantages of the invention will become apparent from the following present detailed description, taken in combination with the appended drawings, in which:

[0043] Figure 1A is a prior art SEM photograph of a single-bus optical micro ring resonator (MRR) in a silicon nitride process with an upper inset photo of the coupling region between the straight and circular waveguide;

[0044] Figure IB is a prior art SEM photograph of a single-bus optical micro ring modulator (MRR) in a silicon process indicating electrical contact pads along with the straight and circular waveguide;

[0045] Figure 2 is a prior art conceptual diagram of a single-bus optical micro ring modulator (MRM) that shows the straight and circular waveguides along with the thermal resistance heater and the PN junction diode inside the ring;

[0046] Figure 3 is a prior art conceptual graph of the optical output response of the MRM device in figure 02 that shows the narrow notch filter behavior of the MRM as a function of wavelength, along with its repetitive cyclical behavior called the free-spectral range (FSR);

[0047] Figure 4 is a prior art conceptual graph of the output response of the MRM device in figure 03 where the notch at 1542-nm has been both greatly magnified and where two values of electrical bias voltage are shown;

[0048] Figure 5A and Figure 5B are a prior art conceptual diagrams of the same single-bus optical micro ring resonator device where (a) shows wider waveguides and (b) shows narrower waveguides as examples of process fabrication variability for the same initial design;

[0049] Figure 6A and Figure 6B are prior art conceptual graphs of the optical output responses for the two figures 5A and 5B, showing the possible change in notch resonance wavelength for the two types of MRR devices;

[0050] Figure 7 is a prior art conceptual diagram of a double-bus optical micro ring modulator (MRM) that shows the thru-straight, circular and drop-straight waveguides along with the thermal resistance heater and the PN junction diode inside the ring;

[0051] Figure 8 is a prior art conceptual graph of the output response of the MRM device in figure 07 where the notch at 1542-nm has been both greatly magnified and where both the through and drop optical output signals are superimposed over each other, but also where the width of the notch / peak is wider than that of the single-bus MRM device in figure 03;

[0052] Figure 9 is a prior art conceptual graph of the output response shown in Figure 8 with two voltages applied to the PN junction where the notch / peak at 1542-nm has been shifted by almost 0.1-nm to the right for the higher voltage;

[0053] Figure 10 is a prior art conceptual graph of the notch of figure 08, where each trace represents the optical output where each 1°C rise in temperature corresponds to a 0.3-nm increase in wavelength;

[0054] Figure 11 is a prior art conceptual graph of three hypothetical locations of notch location at start-up with respect to input laser line wavelength for (a) the notch to the to the left of the laser line, (b) the notch coincident with the laser line, and (c) the notch to the right of the laser line;

[0055] Figure 12 is a conceptual graph showing a hypothetical set of notches spread in groups because of variances in processing and fabrication and highlighting the areas between the clusters of notches where a laser line could be anticipated to be;

[0056] Figure 13 is a conceptual diagram of a double-bus optical micro ring modulator (MRM) that shows the thru-straight, circular and drop-straight waveguides along with the thermal resistance heater and the PN junction diode inside the ring and an optical tap off the through- waveguide directed to a photodetector and an optical tap off the drop-waveguide directed to a photodetector;

[0057] Figure 14 is a conceptual diagram of a double-bus optical micro ring modulator (MRM) similar to figure 13 that shows the drop-waveguide further from the circular waveguide and where no tap is used for the photodetector;

[0058] Figure 15 is a conceptual diagram of a double-bus optical micro ring modulator (MRM) with the differential PID control system for tracking and locking the input optical laser line wavelength to the notches / peaks by regulating the temperature of the device;

[0059] Figure 16 is the conceptual graph of the output optical through and drop ports of the MRM of figure 15, where, at time T=0 (start-up) the notches / peaks of an actively modulated device are initially to the left of the input laser line wavelength. Figure 16B and Figure 16Cshow zoom-in of where the laser line intercepts the curves that in turn generate photocurrents in the two taps to the photodetectors;

[0060] Figure 17 is the conceptual graph of the output optical through and drop ports of the MRM of figure 15, where, at time T=tl (later, tl > 0) the notches / peaks of an actively modulated device have moved right towards the input laser line wavelength. Figure 17B and Figure 17C show zoom-in of where the laser line intercepts the curves that in turn generate photocurrents in the two taps to the photodetectors;

[0061] Figure 18 is the conceptual graph of the output optical through and drop ports of the MRM of figure 15, where, at time T=t2 (later, t2 > tl > 0) the notches / peaks of an actively modulated device have moved right and have locked on the input laser line wavelength for when photocurrents are equal. Figure 18B and Figure 18c show zoom-in of where the laser line intercepts the curves that in turn generate photocurrents in the two taps to the photodetectors;

[0062] Figure 19A is the zoom-in of the notches / peaks showing the position of the notches when the applied high and low PN junction voltages after the notches / peaks have been allowed to shift right under local heating, Figure 19B shows the curve of the error signal and the region where the error signal is negative;

[0063] Figures 20A, 20B and 20C are graphs depicting the dynamics of the simple PID controller when external temperature is changed, and how the error signal as it tends towards zero for normal operation, while error that requires cooling more than what is possible results in a dead-zone of operation;

[0064] Figure 21 is a conceptual diagram of a double-bus optical micro ring modulator (MRM) with the differential PID control system for tracking and locking the input optical laser line wavelength to the notches / peaks by regulating the temperature of the device with the addition of a system to detect the level of un-balanced code in the input electrical modulation signal that affects the gain in the photodetector amplifiers in a complementary manner;

[0065] Figure 22A is a conceptual graph for of the adjustable weighing factors in the photodetector transimpedance amplifier with a linear compensation, and Figure 22b shows a quadratic compensation.

[0066] Figure 23A indicates the lock points in the notches / peaks to the input optical laser line for the case when a linear compensation is used as in figure 22A, and Figure 23B for when a quadratic compensation is used as in figure 22b;

[0067] Figure 24 is a conceptual diagram of a double-bus optical micro ring modulator (MRM) with the differential PID control system for tracking and locking the input optical laser line wavelength to the notches / peaks by regulating the temperature of the device assuming abalanced code, but where the weights are fixed to values that better alter the position of the lock-in point to a more optimum optical amplitude modulation or extinction ratio;

[0068] Figure 25A is a conceptual graph that indicates that the optimized point, where the laser line wavelength exists, is for when the two photocurrents are equal, and Figure 25B shows that for fixed, asymmetric values of weights in the transimpedance amplifiers, the optimized point can be shifted so that a higher OMA is achievable;

[0069] Figure 26A is a portion of the PID control loop specific to the heater resistance and the main driver, Figure 26B is an example of a single-supply amplifier that sources current to the heater resistor, Figure 26C is an example of a dual-supply amplifier that sources and sinks current to and from the heater, Figure 26d indicates that output current of the single-supply amplifier results in no-current for voltages lower than zero, but Figure 26e shows how an absolute-value effect for the current can be achieved with a dual-supply amplifier;

[0070] Figure 27 A, 27B and 27C are graphs depicting the dynamics of the more complex PID controller with absolute value for heater current. Operation follows the same as the simple PID control, except when external temperature pushed the notches / peak into the laser line, and the error becomes negative. The absolute valuing of the current show signal inversion while locking to the alternate set of zero error points in the curves;

[0071] Figure 28A and 28B shows the situation when the input laser line is located within the notches / peaks of the MRM device upon start-up and Figure 27B shows how the dual-supply symmetric amplifier using an effective absolute-value for calculating the error signal allows the notches / peaks to lock to the opposite point where photocurrents are equal;

[0072] Figure 29 shows a circuit that can allow the heater to activate even when the error signal is zero so that the notches / peaks can be “kicked out” of the dead zone at start up and then locked by standard means of the PIC control.DESCRIPTION OF THE INVENTION

[0073] The ensuing description provides representative embodiment(s) only, and is not intended to limit the scope, applicability or configuration of the disclosure. Rather, the ensuing description of the embodiment(s) will provide those skilled in the art with an enabling description for implementing an embodiment or embodiments of the invention. It would be understood by one of skill in the art that various changes can be made in the function and arrangement of elements without departing from the scope of the invention as set forth in the claims. Accordingly, an embodiment is an example or implementation of the inventions andnot the sole implementation. Various appearances of “one embodiment,” “an embodiment” or “some embodiments” do not necessarily all refer to the same embodiments. Although various features of the invention may be described in the context of a single embodiment, the features may also be provided separately or in any suitable combination. Conversely, although the invention may be described herein in the context of separate embodiments for clarity, the invention can also be implemented in a single embodiment or any combination of embodiments.

[0074] Reference in the specification to “one embodiment,” “an embodiment,” “some embodiments” or “other embodiments” means that a particular feature, structure, or characteristic described in connection with the embodiments is included in at least one embodiment, but not necessarily all embodiments, of the invention. The phraseology and terminology employed herein is not to be constmed as limiting but is for descriptive purposes only. It is to be understood that where the claims or specification refer to “a” or “an” element, such reference is not to be constmed as there being only one of that element. It is to be understood that where the specification states that a component feature, stmcture, or characteristic “may,” “might,” “can” or “could” be included, that particular component, feature, stmcture, or characteristic is not required to be included.

[0075] Reference to terms such as “left,” “right,” “top,” “bottom”, “front” and “back” are intended for use in respect to the orientation of the particular feature, stmcture, or element within the figures depicting embodiments of the invention. It would be evident that such directional terminology with respect to the actual use of a device has no specific meaning as the device can be employed in a multiplicity of orientations by the user or users.

[0076] Reference to terms “including,” “comprising,” “consisting” and grammatical variants thereof do not preclude the addition of one or more components, features, steps, integers or groups thereof and that the terms are not to be constmed as specifying components, features, steps or integers. Likewise, the phrase “consisting essentially of,” and grammatical variants thereof, when used herein is not to be constmed as excluding additional components, steps, features integers or groups thereof but rather that the additional features, integers, steps, components or groups thereof do not materially alter the basic and novel characteristics of the claimed composition, device or method. If the specification or claims refer to “an additional” element, that does not preclude there being more than one of the additional elements.

[0077] A “two-dimensional” waveguide, also referred to as a 2D waveguide or a planar waveguide, as used herein may refer to, but is not limited to, an optical waveguide supporting propagation of optical signals within a predetermined wavelength range which guides theoptical signals vertically relative to a substrate upon which the 2D waveguide is formed but does not guide the optical signals laterally relative to the propagation direction of the optical signals within the 2D waveguide.

[0078] A “three-dimensional” waveguide, also referred to as a 3D waveguide, a channel waveguide, or simply waveguide as used herein may refer to, but is not limited to, an optical waveguide supporting propagation of optical signals within a predetermined wavelength range which guides the optical signals vertically relative to the substrate upon which the 3D waveguide is formed as well as laterally relative to the propagation direction of the optical signals within the 3D waveguide.

[0079] A “photonic integrated circuit” (PIC) as used herein may refer to, but is not limited to, the monolithic integration of multiple integrated optics devices into a circuit formed upon a common substrate providing an optical routing and processing functionality. The PIC is fabricated using processing techniques at a wafer level, e.g. CMOS manufacturing flows, MEMS processing flows, etc.

[0080] A “wavelength band” as used herein may refer to, but is not limited to, a defined wavelength range over which a device or system is either intended to operate or provide functionality. Such wavelength bands may be associated with a standard, such as a telecommunication standard for example, by the design of the optical component(s) / system(s) or by an informal reference as known to those of skill in the art. A wavelength band may refer to the “O-band” which as used herein refers to, but is not limited to, the wavelength range 1260-1360nm. A wavelength band may refer to the “E-band” which as used herein refers to, but is not limited to, the wavelength range 1360-1460nm. A wavelength band may refer to the “C-band” which as used herein refers to, but is not limited to, the wavelength range 1530-1565 nm. A wavelength band may refer to the “S-band” which as used herein refers to, but is not limited to, the wavelength range 1460-1530 nm. A wavelength band may refer to the “L-band” which as used herein refers to, but is not limited to, the wavelength range 1565-1625 nm. A wavelength band may refer to the “U-band” which as used herein refers to, but is not limited to, the wavelength range 1625-1675nm.

[0081] Embodiments of the invention may be “hybridly integrated.” This may, within some embodiments of the invention, refer to, but not be limited to, the “integration” of an optical element onto a substrate by attaching the optical element or another element physically integrated with the optical element to the substrate (platform) such that the optical element is retained in position. Such attachment means may include, but not be limited to, soldering, epoxy, van der Waals forces, electrostatic attachment, magnetic attachment, physicalinterlocking and friction. Accordingly, in these embodiments of the invention the optical element being hybridly integrated may be viewed as being implemented within a parallel manufacturing process to the other optical element(s) prior to being co-assembled. This parallel manufacturing process may employ one or more processes selected from the group comprising, but not limited to, liquid phase epitaxy (LPE), metal organic chemical vapor deposition (MOCVD), organometallic vapor-phase epitaxy (OMVPE), selective area epitaxy, an additive manufacturing process, a non-additive manufacturing process, crystal growth, doping, induced damage, etching, doping and deposition.

[0082] This may, within other embodiments of the invention, refer to, but not be limited to, the “integration” of an optical element onto a substrate using a different manufacturing methodology and / or techniques to those employed in forming other optical components upon the substrate. For example, this may employ an LPE process to form the other optical element upon the substate wherein the optical component upon the substrate was formed by MOCVD or vice-versa. Alternatively, both the optical component and other optical component may be formed using the same manufacturing methodology or a combination of manufacturing methodologies. These manufacturing methodologies may employ one or more processes selected from the group comprising, but not limited to, LPE, MOCVD, OMVPE, selective area epitaxy, an additive manufacturing process, a non-additive manufacturing process, crystal growth, doping, induced damage, etching, doping, deposition, an additive manufacturing process and a non-additive manufacturing process. Accordingly, in these embodiments of the invention the optical element being hybridly integrated may be viewed as being implemented within one or more further processing stages of the same manufacturing process as the other optical element(s). However, in each instance the optical waveguide and / or optical component properties require that an optical interface is implemented between the optical waveguide and optical component in order to provide efficient optical coupling between one and the other.

[0083] Embodiments of the invention may be “monolithically integrated.” This may, within some embodiments of the invention, refer to, but not be limited to, the “integration” of optical elements onto or within a substrate directly forming each optical component onto or within the substrate (platform). The manufacturing processes for the optical elements may be concurrent, plesiochronous or asynchronous. Each optical element may be formed from one or more processes selected from the group comprising, but not limited to, LPE, MOCVD, OMVPE, selective area epitaxy, an additive manufacturing process, a non-additive manufacturing process, crystal growth, doping, induced damage, etching, doping and deposition.

[0084] An optical element may employ one or more semiconductors grown using LPE, MOCVD, and OMVPE, for example. The one or more semiconductors may be selected from, but not limited to, group III-V semiconductors, group II-VI semiconductors, group IV semiconductors, and group IV-V-VI semiconductors. Examples of group III-V semiconductors may include A1P, AIN, AlGaSb, AlGaAs, AlGalnP, AlGaN, AlGaP, GaSb, GaAsP, GaAs, GaN, GaP, InAlAs, InAlP, InSb, InGaSb, InGaN, GalnAlAs, GalnAlN, GalnAsN, GalnAsP, GalnAs, GalnP, InN, InP, InAs, InAsSb, and AllnN. Examples of group II-VI semiconductors may include ZnSe, HgCdTe, ZnO, ZnS, and CdO. Examples of group IV Semiconductors may include Si, Ge, and strained silicon. A group IV-V-VI semiconductor may be GeSbTe.

[0085] Within embodiments of the invention the platform or substrate upon which the integration is performed may be a silicon substrate wherein the one or more optical waveguides upon the platform exploit a silicon nitride core with silicon oxide upper and lower cladding, a SiO2— Si3N4— SiO2waveguide structure. Alternatively, the one or more optical waveguides may employ a silicon core with silicon nitride upper and lower claddings. Optionally, the upper cladding may be omitted within other embodiments of the invention.

[0086] However, it would be evident that other optical waveguide structures may be employed including, but not limited to, silica-on-silicon, doped (e.g., germanium, Ge) silica core with undoped cladding, silicon oxynitride, polymer-on-silicon, or doped silicon waveguides for example. Additionally, other waveguide structures may be employed including vertical and / or lateral waveguide tapers and forming microball lenses on the ends of the waveguides via laser and / or arc melting of the waveguide tip.

[0087] Further, whilst embodiments of the invention are described with respect to silicon-on- insulator (SOI) waveguides by way of example, e.g. SiO2— Si3N4— SiO2; SiO2— Ge SiO2— SiO2; or Si — SiO2„ it would be evident that within other embodiments of the invention may be employed to coupled passive waveguides to active semiconductor waveguides, such as indium phosphide (InP) or gallium arsenide (GaAs), e.g. a semiconductor optical amplifier (SOA), laser diode, etc. Optionally, an active semiconductor structure may be epitaxially grown onto a silicon IO-MEMS structure, epitaxially lifted off from a wafer and bonded to a silicon integrated optical microelectromechanical systems (IO-MEMS) structure, etc.

[0088] However, within other embodiments of the invention a variety of waveguide coupling structures coupling onto and / or from waveguides employing material systems that include, but not limited to, SiO2— Si3N4— SiO2SiO2— Ge: SiO2— SiO2Si — SiO2ion exchanged glass, ion implanted glass, polymeric waveguides, InGaAsP, GaAs, III-V materials, II-VImaterials and optical fiber. Whilst primarily waveguide-waveguide systems have been described it would be evident to one skilled in the art that embodiments of the invention may be employed in aligning intermediate coupling optics, e.g., ball lenses, spherical lenses, graded refractive index (GRIN) lenses, etc. for free-space coupling into and / or from a waveguide device.

[0089] Further, whilst embodiments of the invention are described primarily with respect to a silicon substrate it would be evident that other substrates may be employed within other embodiments of the invention. These may include, but not be limited to, a semiconductor, a ceramic, a metal, an alloy, a glass, or a polymer.

[0090] A “ceramic” as used herein may refer to, but is not limited to, an inorganic, nonmetallic solid material comprising metal, non-metal or metalloid atoms primarily held in ionic and covalent bonds. Such ceramics may be crystalline materials such as oxide, nitride or carbide materials, elements such as carbon or silicon, and non-crystalline. Exemplary ceramics may include high temperature ceramics or high temperature co-fired ceramics such as alumina (A12O3), zirconia (ZrO2), and aluminum nitride (AIN) or a low temperature cofired ceramic (LTCC). A LTCC may be formed from a glass - ceramic combination.

[0091] A “glass” as used herein may refer to, but is not limited to, a non-crystalline amorphous solid. A glass may be fused quartz, silica, a soda-lime glass, a borosilicate glass, a lead glass, an aluminosilicate glass for example. A glass may include other inorganic and organic materials including metals, aluminates, phosphates, borates, chalcogenides, fluorides, germanates (glasses based on GeO2), tellurites (glasses based on TeO2), antimonates (glasses based on Sb2O3), arsenates (glasses based on As2O3), titanates (glasses based on TiO2), tantalates (glasses based on Ta2O5), nitrates, carbonates, plastics, and an acrylic.

[0092] Further, whilst the embodiments of the invention may be described and depicted with respect to a specific waveguide geometry, for example one employing a core and lower cladding, a so-called rib waveguide, it would be evident that other waveguide geometries such as a buried waveguide, diffused waveguide, ridge or wire waveguide, strip-loaded waveguide, slot waveguide, and anti-resonant reflecting optical waveguide (ARROW waveguide), photonic crystal waveguide, suspended waveguide, alternating layer stack geometries, subwavelength grating (SWG) waveguides and augmented waveguides (e.g. St — SiO2— Polymer) may be employed without departing from the scope of the invention. Further, these waveguide geometries may be step index, graded index or hybrid index (combining inversestep index and graded index).

[0093] Within the following description elements within the figures are identified by providing their reference numeral within a pair of square brackets, e.g.

[1234] ,

[0094] A “photodetector” as used herein may refer to, but is not limited to, a sensor of light or other electromagnetic radiation. Photodetectors may be classified by their mechanism for detection which may include, but not be limited to, the photoconductive effect, photoelectric effect, thermal, polarization, and photochemical: Photons cause electrons to transition to midgap states then decay back to lower bands, inducing phonon generation and thus heat. Semiconductor-based photodetectors typically use a p-n junction that convert photons into charge and may be single junction designs or multiple junction designs such as an avalanche photodiode (APD). A photodetector as used herein may include detectors exploiting surface plasmon or other surface-state carrier generation effects.

[0095] Overview.

[0096] Optical devices, in particular those that use some type of active or dynamic electrical charge or field that is associated with some form of electron carrier density can be sensitive to the temperature within the structure, the so-called junction temperature, along with the ambient temperature outside the device. This is especially true for devices such as the optical Micro Ring Modulator (MRM), also known as micro-ring resonators (MRR) for when no active modulation is present and used only as a passive optical structure such as a wavelength filter.

[0097] Figure 1A shows a prior art 3D SEM photograph of a simple MRR device, where the straight waveguide is tangent to a circular waveguide. Light in the range of 1550-nm wavelength is coupled into the cores of these guides and complex coupling and interference effects are observed. The Figure IB shows a prior art 3D SEM photograph of a more complex MRM device where the initial waveguides are present but they are surrounded by other elements such as a heater resistance and a PN junction modulation structure as well.

[0098] These devices are typically tuned to use very narrow laser linewidths and use constructive and destructive interference to modulate light. These devices do suffer from the effects of thermal drift that can lead to large changes in the position of their spectral peak / notch relative to the input wavelength of light causing poor or no functionality.

[0099] An MRM device that uses the PN junction to modulate a signal typically has the middle of the active region (where the P and N meet) within the center of waveguide itself. This way, the light in the ring can pass through the center region of the PN junction. By changing the bias conditions on the PN junction (by either forward or reverse biasing voltage), the number of free-charge carriers can be changed in this region. The amount of free charge in the region dictates the effective index of refraction in this area and thus the speed of the light in that partof the waveguide. When there is excess charge in the region, the effective index of refraction increases and light tends to travel slower, and when the charge is removed from this region the effective index of refraction decreases and the light tends to go faster. By varying the speed of the light in the circular waveguide, the effective distance in the circle is changed; although physically the diameter of the ring really does not actually change, such that the optical path length is increased or decreased as the refractive index of the waveguide is changed. As the light travels around the circle, back to the coupling region with the straight guide, if it has been lagged (or leaded) with respect to the initial light, this produces the effect of constructive or destructive interference, which in turn results in an amplitude change in output light intensity at the through-port.

[0100] As light travels roughly 2xl08m / s in the Silicon waveguide, and the ring itself is only 10’s of microns in diameter, light dwells in the ring system for many billions of billions of wavelengths. This is a measure of the photon lifetime in the ring and leads to performance dynamics such as the Q-factor of the ring which is a measure of the bandwidth of the filter - or how quickly it reacts to inputs, similar to the responses of RLC or LC circuits in electronics. A high value of Q means that the filter can be highly specific selecting a wavelength (a narrow notch), but it typically cannot accommodate fast changes in the input signal. However, a lower value of Q means that the notch is less pronounced, allowing more wavelengths to pass, but its position in wavelength can be adjusted faster. It is an operational design compromise in many designs.

[0101] In MRM devices that use the semiconductor PN-junction diode for modulation, temperature plays a major role in performance. Since the device uses interference effects, it is very dependent on slight changes to the carrier densities in the valence and conduction bands that are in turn also very dependent on the temperature of the system.

[0102] Although temperature effects have been proven to be fairly easy to regulate in standard electronics with proper circuit design, thermal management and initial process control of the doping profiles of the devices. In the case of silicon photonic MRM devices that use the PN junction, not only are the general physical structures much bigger and have larger, irregular geometries, but the method of using interference techniques to create amplitude changes using constructive or destructive interference to modulate output power means that the sensitivity to small variations such as temperature is amplified dramatically. In fact, so much so that MRR and MRM devices have often been used to measure small variations in temperature using broad spectrum light sources and tracking how nulls in the spectrum move.

[0103] Basic Micro Ring Modulator:

[0104] A simplified silicon waveguide optical Micro Ring Modulator (MRM) device structure is shown in Figure 2 and shows the input straight waveguide

[0001] , the circular waveguide

[0003] that is close but not touching the straight waveguide, along with representations of the thermal tuning element resistor

[0005] and the PN junction diode element

[0007] . The input optical port

[0009] and the output through port

[0011] are shown as well as the voltage stimuli for modulating the PN junction

[0013] and applying voltage to the thermal resistor

[0015] . Note that this figure is descriptive and that the physical implementation of these structures are typically in 3 dimensions.

[0105] This type of MRM device is called a single-bus optical micro ring modulator and assumes that at least a portion of the device is made from semiconductor material that can allow for carrier mobility in the PN junction portion. In the case where no modulation signal is applied (i.e.: no PN junctions are formed), the devices can be made from other materials such as Silicon Nitride, where it typically functions as more of an optical filtering device particularly for filtering wavelengths but where temperature still plays a large role.

[0106] The MRM is a form of spectral notch filter, it “removes” a particular wavelength from a set or continuum of wavelengths. It has a particular resonant frequency (wavelength) that causes destructive interference that removes a wavelength. In this case, if the input optical port of the device

[0009] receives narrow linewidth wavelength from a tunable laser light source and this wavelength is tuned from (in the graph) 1534-nm to 1566-nm, the output optical port of the device

[0011] responds with the characteristic output resonance response represented in Figure 3. The example shows a series of attenuations of optical power, the first center notch located at 1542 -nm wavelength

[0017] , with the next at 1552-nm, a full 10-nm separation in its repeating free-spectral range (FSR)

[0019] pattern. Effectively, the FSR is related to the occurrences where the wavelength of the light fits into the diameter of the ring an integer number of times. The width of the notch is related to weak-coupling ratios between the straight and circular guides and directly affects the Q-factor, a measure of the narrowness and steepness of the curve. In this case the full-width half maximum (FWHM) of the notch is roughly 70- picometers

[0021] ,

[0107] For the same device, but focusing in on the 1552-nm notch, two different reversebias voltages on the PN junction diode are now shown in Figure 4. Note that forward-biasing the PN junction can also have the effect of charge displacement, but we will assume reversebias as this method can have a lower static power consumption.

[0108] In this case, the first curve

[0023] is for a 0-V bias and the second curve

[0025] is for a 10-V reverse-bias that moves the notch to a wavelength of 1552.08-nm. The difference innormalized amplitude is roughly 0.7

[0027] at the fixed laser wavelength applied at 1552-nm. The PN junction voltage of course can be modulated at very high speeds compared to the thermal changes, but it is limited in total wavelength displacement of the solid and dashed curves to only 10’ s of picometers for even very large voltages. However, if the notches are pronounced enough, with a high Q-factor, a very large change in output optical power can be achieved during modulation.

[0109] Optical transmission systems typically use high and low levels of optical power and not the “rail to rail” signals typical of electronics. The data stream must therefore be composed of a largely balanced train of 0’ s and 1 ’ s such that a moving window average of the data appears to have a roughly 50% average optical power between its optically ‘high’ power and its optically ‘low’ power. To do this, a standard stream of 0’s and l’s (e.g.: a set of bytes from a classical computer calculation) is re-arranged using balancing codes, common examples are called 8B / 10B or 64B / 66B encoding algorithms, which map regular 8-bit bytes to longer groups of 10-bits that are unique but have almost the same number of 0’s as l’s in their codes. The reason to maintain this balanced average is because most optical signaling methods do not have an absolute “off’ level where the optical power completely goes to zero (in many cases, completely tuming-off the light signal can degrade the speed of the modulation and is related to the dynamics of either the laser source itself or the natural extinction ratio of the modulator). Therefore, typically a photo-detector receiver circuit receives a “high” and “low” signal and not an “on” and “off’ signal. This is also helpful because regardless of how far the signal has travelled and has become attenuated, the average of the signal is always between the high and the low states and so the data sampled relative to its average is either “above the average” for a logical one, or “below the average” for a logical zero. Tracking this average and then sampling above or below it typically involves some form of phase-lock loop design like automatic gain control (AGC), and comparators, but it is an exceptionally robust way to transfer data.Irrespective of how thermal control is implemented, balanced code plays an important role in controlling the MRM device as well. In prior-art implementations of thermal control, light is sampled from the output of the single-bus MRM, typically using an optical splitter and a low- speed photodetector. With the signal being optically balanced, a slowly varying time-average of this monitor signal can be made that is data independent and is related to the position of the notch with respect to the laser wavelength where the greatest magnitude corresponds to the greatest change in low-to-high signal (and hence the center of the notch). This signal is then typically sampled using an analog-to-digital converter (ADC) where it is stored as a referencesignal. The active signal is then DSP processed using an FPGA chip that generates another slowly varying digital signal that can be used in a comparison circuit.

[0110] Fundamental Tuning Problem:

[0111] Therefore, the main obstacle for the practical use of the MRM is its thermal stability in wavelength. This instability arises from mainly 2 sources: 1) process fabrication errors and 2) carrier density related to temperature (especially for silicon PN junctions). While the first is error is static and due to process control of today’s silicon photonics processing methods such as etching and doping, the second is dynamic and continuous. In either case, the desired location of the notch wavelength will not be where it was designed to be. And an active control tracking and locking systems is needed.

[0112] In Figure 5A the single bus MRM device is shown as the ideal design case and its subsequent spectra is shown in Figure 6A. Figure 5B shows how the diameter of the circle and the bus might be over-etched resulting in narrower waveguides with the resulting spectra shown in Figure 6B, and its displaced notch relative to Figure 6A. This is only a very simplified comparison, and other effects could also ensue.

[0113] With regards to temperature dependence, the PN-junction diode structure’s is well studied and relates the valence and conduction band carrier density dependence to temperature. While normal electronic applications are not as affected by temperature, the MRM device is based on interference effects and therefore small changes to carrier density can have large optical spectrum effects.

[0114] Therefore, while an initial design may specify the notch to be placed at 1550-nm, any deviations in etching, doping and thermal dissipation may shift the real notch several nanometers from the ideal position. This makes aligning the laser wavelength more challenging - as the notch needs to be located first. But, as the device increases and decreases in temperature, the notch naturally moves as well. In addition, using the same narrow linewidth input light source for multiple MRM devices on the same chip becomes even more difficult because each MRM device may have different positions for their own notch wavelengths for the same temperature.

[0115] Double Bus Optical Micro Ring Modulator:

[0116] A second type of MRM device is shown in Figure 7 and shows a secondary straight waveguide

[0029] , this is called a double-bus optical ring resonator. The double bus MRM is a four port device, with the straight waveguide on the top

[0031] called the through-bus and the straight waveguide on the bottom

[0029] called the drop-bus. The ports are now labeled as the constant light input port

[0033] , the through port

[0035] , the drop port

[0037] , and the add port

[0039] .The device still works appreciably in the same way as the single-bus device, except that light can now be coupled out the through port or the drop-port depending on thermal and electrical settings. Furthermore, an injected signal from the add-port can also be channeled out the thru- port making the 4 port device act as a 2x2 switch. The thermal tuning resistor

[0041] (the dropbus waveguide is drawn under the thermal tuning resistor and does not divide it) and the applied voltage on the PN junction

[0043] are again used to dynamically change the state of the device.

[0117] A subtle difference between the single and double MRM devices is that normally, because of the added coupling losses from the second coupling region

[0045] , the through and drop port powers tend to exhibit wider or more broad transfer functions (lower Q-factors) as shown in Figure 8 for the through

[0047] and for the drop

[0049] optical power versus wavelength, with the drop even wider than the through port.

[0118] Although a narrow notch, or High-Q factor, does allow for high selectivity, because of the longer photon lifetime this type of device, it is normally more difficult to modulate at higher speeds as the photons take time to dissipate and accumulate. With a well-designed double-bus MRM, its wider notch, or lower-Q (but not much lower), can produce a high modulation rate. However, some care in comparing the single and double -bus data rates must be observed. If we assume the magnitude of the voltage from the driver attached to the PN junction diode is the same for the single and double bus configurations, then, a 10-V swing would result in a lower optical amplitude for the double-bus than for the single-bus due to the less steep curves. Figure 9 shows the through

[0051] and drop

[0053] curves for 0-V and the through

[0055] and drop

[0057] curves for 10-V with the OMA

[0059] at the laser line of 1552-nm just as before. If we relate this to the Single-Bus MRM of Figure 3 with the same voltage change causing the same displacement in wavelength, a lower relative amplitude change with a value of roughly 0.54 is observed. To have the same OMA, a stronger driver would be needed, but also the ability to push the MRM device itself to a higher, red-shifted amount in wavelength.

[0119] Restrictions on Temperature Control:

[0120] In the following descriptions of thermal tuning and tracking using a novel method for control, the assumption is that the control system can only be used to actively HEAT the device using a thermal resistor made using a metal like tungsten located around and near to the MRM device. By HEATING the device, the notch always tends towards a red-shifting, in other words towards the longer wavelengths towards the right of the graph. In a silicon process, the amount by which an MRM device can be shifted towards the right of the graph is roughly 0.07-nm / °C as shown in Figure 10 from a Temperature of 22°C at

[0061] to almost 35°C at

[0063] and the arrow

[0065] indicating further red-shifting to the right.

[0121] The material processing and fabrication of the MRM device (typically in silicon), available at present, does not have the device structures to actively COOL individual devices, although chip-scale Peltier TEC coolers could be employed but that would be applied to all devices on a chip at the same time. If at some point in the future a localized active cooling element becomes possible to fabricate near the MRM, generally, the method described herein could be implemented in all cases and there would be no conditions on operation (except for obvious design maximums of operation).

[0122] This technique then requires that the system allows for a thermal time-constant that can passively cool the device and allow for the notch wavelength to drift back towards a blue- shifted and shorter wavelength toward the left of the graph.

[0123] The ACTIVE-HEATING and PASSIVE-COOLING method has one significant drawback. If the narrow linewidth source laser light

[0067] ends up, upon start-up, inside

[0069] or to the left

[0071] (shorter wavelength) of the notch as shown in Figure 11 A Figure 11 B and Figure 11c, the controller has no active cooling and therefore the notches cannot be actively blue- shifted to shorter wavelengths (on a per device basis).

[0124] However, with all the design parameters available such as input laser wavelength

[0073] , device geometries and sufficiently accurate simulation along with the expected spectral distance of the FSR resonance peaks (~10-nm)

[0075] and the narrowness of the notches (~100- pm), it is likely that with a very large probability, the MRM devices can be designed such that, within all errors already described resulting in a random position of a cluster of notches

[0077] , the laser linewidth can be located in areas to the right of the notches

[0079] at the initial temperature as shown in Figure 12. If the MRM device initially requires active heating to move the notch(es) to longer wavelengths, then passive cooling can then be used to help lock and track the notches to the laser wavelength after that. This is analogous to using the gas pedal of a car to get up to a constant speed, but rather than use braking to slow the car (should it go a little faster), simply let up on the gas pedal and allow the car to slow itself.

[0125] However, some techniques will be described later to avoid initial start “dead-zones” after the main method for the control system is described, this will address the issues of when the laser linewidth is inside or to the left of the notches initially.

[0126] Differentially Tuned Double-Bus Optical Ring Resonator:

[0127] The design shown in Figure 13 shows a double-bus MRM with additional tap monitors that detects a portion of the output power on the through

[0081] and drop

[0083] ports. The first photodetector

[0085] , called PDmonl is used to capture a portion of the optical power passing out the through-port using the first optical splitter

[0087] (typically a multi-modeinterferometer (MMI)) and generates a photo-current of IpDmoni

[0089] . The second photodetector

[0091] , called PDmon2 is used to capture a portion of the optical power passing out the drop-port using the second optical splitter

[0093] (also typically a multi-mode interferometer (MMI)) and generates a photo-current of IpDmon2

[0095] . The ratio for the MMI splitters can be tailored to any desired optical power from 50 / 50 to 90 / 10 or others. Of course, the less tapped power the better it is for the transmitted output signal (especially when used as an optical transceiver) but the amount of tapped power must still maintain above an acceptable noise level.

[0128] In the case of an optical transceiver application, where the primary goal is to couple optically modulated signals into an optical fiber, using the double-bus with a “drop-port” may not be optimal as it may lower the overall output power and may cause slower modulation rates. However, tuning the ring to allow for asymmetric coupling ratios at the weakly coupled regions where the buses meet the ring, can allow for the drop port to have much less light. Figure 14 shows a similar Double-Bus Optical Ring Resonator where the secondary bus has a lower coupling ratio

[0097] by essentially making it further away from the ring, which is then directed into a photodetector without a splitter. The much lower power drawn out of the circular waveguide into the drop port can then be completely directed into PDmon2 without a splitter.

[0129] While many types of structures can be designed, the following descriptions will assume the design of Figure 13 as the more generic design structure.

[0130] Controller for Thermal Locking and Tracking:

[0131] A simplified electrical / optical circuit diagram for the control system is shown in Figure 15 and shows all the elements as previously described along with the PID (Proportional Integral Derivative) Controller system linking the differential photodetector currents to the heater near the MRM device. The transient behavior of the Control Theory of PID controllers will not be described here such as the Gain or the specific filter poles that are required to properly dampen the system (i.e.: the I and D portion of the design), as this is very dependent on the materials, physical geometries and the overall layout. However, it is important to properly design all aspects of transfer function and feedback to keep the controller stable.

[0132] The additional aspects included in the figure are the input stimulus

[0099] to the PN junction

[0101] . The complementary photodetector currents

[0103]

[0105] , that are amplified and averaged by the Transimpedance and averaging circuits

[0107]

[0109] . These average voltages

[0111]

[0113] are subtracted by the summer circuit

[0115] and the error signal

[0117] is sent to the PID control

[0119] that sources current to the thermal resistor

[0121] ,

[0133] First note that the photodetector currents are complementary, and as one increases the other decreases. Also, the 0’s and l’s bit pattern

[0099] use balanced codes and are modulatedat least an order of magnitude faster than the fastest thermal transient for the device (i.e.: bit rate is faster than the fastest that the device can heat or cool). Lastly, it is assumed that the device is started at a lowest set environmental temperature (for example room temperature) and that the MRM device has been designed so that the input laser line will be on the right side of the notches as in Figure 11A (i.e.: slightly lower in wavelength).

[0134] The objective of a PID controller is to trend the error-signal <IoiFFave>

[0117] (the subtraction of two measured average signals <IpDmoni>

[0111] and <IpDmon2>

[0113] ) towards zero. This can be done between a reference value and a measured value, or in this case, two measured values that are complementary.

[0135] At time T = 0, Figure 16A shows the laser line to the right of the notches and it produces a much greater average current from <IpDmoni>

[0123] than <IpDmon2>

[0125] as can be seen in the zoom in Figure 16B and Figure 16C. The error signal generated is therefore high and causes the heater to tum-on rapidly (again, proper damping and gain for the controller must be designed). This starts to shift the notches towards the laser line as the MRM device heats up.

[0136] At time T = tl, Figure 17A shows the laser line still to the right of the notches and still producing a greater average current from <IpDmoni>

[0127] than <IpDmon2>

[0129] , but the magnitude of the error signal is not as strong. Figure 17B and Figure 17C provide a zoom in again of the monitor currents and their average values.

[0137] At still a later time T = t2, Figure 18A now shows the condition where the laser line and the position of the notches make both average currents <IpDmoni>

[0131] and <IpDmon2>

[0133] equal and where the magnitude of the error signal has decreased to zero. Figure 18B and Figure 18c show zoom in views to the monitor currents as well.

[0138] Figure 19A shows a close-up of the thru and drop ports at a 0-V bias

[0135]

[0137] and the thru and drop ports with a 10-V bias

[0139]

[0141] , The error signal shown in Figure 19B is generated by the subtraction of the window averages, of a period T, of the photocurrents:1 t+T 1 t+TErrorsignai= -ftIPDmonl(u)du — - ftIPDmon2(u)du

[0139] There are two points in this graph where the error-signal is zero; at

[0149] , and at

[0151] . These two points are unique but are also dependent on the magnitude of the voltage change applied to the MRM device. The zero point

[0147] is the point for which the PID controller in its normal mode of operation described above will be locked. The symmetric point

[0149] can also be used to lock the PID controller, however the “heating only” strategy then requires an alternate formulation which will be described later.

[0140] Note that this symmetry depends on the physical design of the MRM in that certain MRM structures may result in asymmetric Lorentzian curves for through and drop outputs. If this occurs, the condition that <IpDmoni> = <IpDmon2> cannot be uniquely satisfied and a slight oscillatory behavior between <IpDmoni> = <IpDmon2> + d / 2 and <IpDmoni> + d / 2 = <IpDmon2> exists. The small where delta ‘d’ value is the minimum value between two wavelengths where this is satisfied. In other words, the error signal never settles exactly to zero, it oscillates between two values, one above zero and one below zero each corresponding to two wavelengths. Although this does not necessarily mean proper performance cannot still be achieved, with proper design considerations for the MRM structure this can be made close to negligible.

[0141] The dynamics of the simple PID control is shown in Figures 20A 20B and 20C. These graphs show (a) the external and local temperatures, (b) the voltages of the error and heater, and (c) the net power provided to the heater. To be clear, for a set tracking of a specific temperature, as long as some minor heating and passive-cooling can be allowed, the control will remain locked indefinitely. However, as shown in Figure 20A, if the external temperature is allowed to climb to a point where the red-shifting requires the laser line to exist inside the notches / peaks, this system will lose-lock as a dead-zone will occur as indicated by the region between 0.010s and 0.015s. Figure 20B shows that during locking, even if the temperature is increased, the error signal will tend to zero. However, if the external temperature exceeds the ability to red-shift the notches / peaks, the error signal will become increasingly negative and the heater voltage will tend to zero - in other words, the MRM device is now too hot and cannot be cooled. Lastly, Figure 20C shows specifically the amount of power delivered to the heater and how this tends to zero with the dead-zone.It should be noted that while this tracking method explicitly shows how a set of notches / peaks can be shifted in wavelength to lock to a static laser line, this same technique can also be used to track to a moving laser line. In other words, if the laser line itself is moved to higher or lower wavelengths, the MRM and PID control would be able servo-lock to a laser line that did not have a fixed wavelength. This PID control can help in three ways: 1) counter-balance any error in processing, 2) counter-balance any environmental changes in temperature and 3) maintain a lock to an input laser wavelength that is slowly varying in wavelength.

[0142] Controller for Thermal Locking and Tracking with Un-balanced Code:

[0143] While the majority of use cases for the MRM device will likely be for optical transmission of very high speed data, there are other examples of data transmission and processing that might use strictly un-balanced, or raw, code. Examples of this are burst-mode data transmission and digital logical functions, respectively.

[0144] The slightly modified electrical / optical circuit shown in Figure 21 shows all substantially the same components described in Figure 15 above, with the addition of a normalized moving windowed average process

[0155] (also called a low-pass filter in the frequency domain) carried-out directly on the high-speed modulation signal

[0157] . Where the electrical modulation data stream of 0’s and l’s is now completely unconstrained and can be un-balanced (or balanced) representing any set of digital byte sequences.

[0145] The normalized windowed average signal is generated that represents the level to which the code is balanced; from 0 (meaning all 0’s) to 0.5 (meaning perfect 50% balancing) to 1.0 (meaning all l’s). This signal can be used to change the gain of the TIA’s

[0159]

[0161] from each of the photodetectors in a complementary way.

[0146] The two variable-gain TIA and Averaging circuits

[0159]

[0161] still would produce average photocurrent signals, only that they are now weighted by a factor of a or b for produce <aIpDmoni> and <bIpDmoni>

[0163]

[0165] . As it is still a PIC control circuit, and the error still needs to track towards zero, the weighting factor could (for example) increase the signal <IpDmoni> if it was tending to see more optical zeros by a factor of a, and decrease the signal <IpDmon2> if it was tending to see more optical ones by a factor of b. Eikewise, the opposite weighting could occur if <IpDmoni> was higher because it had more ones and <IpDmon2> was lower because it had more zeros.

[0147] There are several algorithms that can be used to calculate the a and b weights, two examples are shown in Figure 22A and Figure 22b. The linear compensation

[0167] is the easiest of all algorithms requiring little extra circuitry. The Gain and bias of the linear curves needs to be tailored to the specific MRM device as it becomes part of the overall loop-gain and frequency response of the PID control. As shown in Figure 23A, the exact alignment of the notches to the laser line can vary depending on the amount of balancing in the code. This may be appropriate for some applications that are expected to remain thermally stable for long periods of time. But does lack some precision.

[0148] A more precise algorithm is based on a quadratic fitting of the best case weights for achieving much lower deviation in lock point. The quadratic coefficients (y = Ax2+BX + C)

[0169] for the curves shown in Figure 22b must be calculated from the performance of the specific type of MRM designed, but once known can offer a very stable and repeatableweighting strategy for all similar devices. Of course, this does come at the expense of more complicated circuitry. Figure 23B shows an example of how well the lock in point is held for vastly different code balancing.

[0149] Shifting Lock-Point closer to Optimal OMA:

[0150] Applications exist such as long-distance optical transceiver technologies, where light levels become very attenuated over long distances of single mode fiber, and thus optimal optical modulation amplitude (OMA) is desired. Also, for even more sensitive optical links that involve optical amplification (such as Erbium or Raman amplification) where signal and noise components can both be amplified, optimal optical extinction ratio (Er) is even more important. A control system is desired that forces the location of the laser line with respect to the notch and peak of the thru and drop ports to be coincident with the lowest point on the thru notch or drop notch.

[0151] The afore described PID method primarily obtains a lock point where the thru notch intersects the drop peak, for a high bias on the PN junction or the where the thru peak intersects the drop notch for a low bias on the PN junction. This is usually some 10’s of picometers from the optimal OMA at the base of the thru notched (or drop notch).

[0152] However, this circuit can be modified in a similar manner to the non-balanced code, but instead using balanced-code where the weights of the two Variable-Gain TIA and averaging blocks

[0159]

[0161] are made to be permanently unbalanced. For example, in Figure 24, if the a coefficient is set to 0.5

[0173] and the b coefficient is set to 1.5

[0175] , the skewed Gain

[0177] and

[0179] would make it appear that <IpDmoni> is weaker then <IpDmon2>, even though they would be the condition for equality, the a and b weights would cause the system to slightly bias the lock point towards a higher OMA, i.e.: closer to the lowest part of the notch. Figure 25A again shows the balanced code for equal photocurrents where a = b. Figure 25B shows the case where a is less than 1 (for example 0.5) and b is greater than 1 (for example 1.5). The laser line

[0181] in both cases is at the same place, but the previously set position for equality

[0183] has been slightly moved towards the right aligning the laser line closer to the optimal OMA point.

[0153] These coefficients must be chosen carefully and in line with the characteristics of the MRM device so that unwanted feedback, gain or oscillation occur.

[0154] Dead-Zone Adaptations for PID Controller:

[0155] The earlier descriptions of the PID controller assumed that the laser line exists to the right of the notches / peaks. Therefore, a simple initial heating of the controller moves the notches / peaks towards the laser line, and the active-heating / passive-cooling process can maintain lock indefinitely after that as long as external temperature remains fairly constant.

[0156] In addition, if the laser line happens to be located just to the left of the notches / peaks upon start up, the PID controller described above will attempt to heat the MRM device until it can push the entire FSR one complete period and red shift the spectrum until it can lock on the next available set of notches / peaks. This is a solution that may be power hungry, and the physical compliance of the structures to the physical requirements of the currents and temperatures may be difficult to satisfy without damaging the structure. However, it should be noted that this technique, if planned well for thermal loading and power, could be a way to establish a control system that could compensate for very large external environmental conditions such as the military specifications for -40C to +125C operating temperatures.

[0157] However, an alternate PID driver system can be envisioned that can eliminate the start-up dead-zone - where the laser line initially aligns with, or is forced to align with (by increasing external temperature for example) the notches / peaks of the MRM.

[0158] Figure 26A shows only the last portion of the PID control circuit as seen before. Figure 26B shows an example of the simple PID control circuit where a last stage amplifier drives the final heating resistor. In this case, if the applied voltage VIN (a figurative composite voltage resulting from the PID control), becomes negative, the current from the driver must go to zero as shown in the earlier graphs in Figure 20B and 20C. Because the simple PIC controller amplifier assumes a single-rail power supply shown in Figure 26B. However, if the amplifier circuit can use both positive and negative supply rails as in Figure 26C, then the current output is effectively absolute valued since the power delivered by a resistor is P = I2R, and the current can now be either sourced or sunk to / from the resistor. This allows a negative error signal to be used as shown in Figures 27 A, 27B and 27 with the dynamics of this improved PID controller. During the time interval between 0 and 0.010s, the improved PID controller acts in the same way as the simple PID controller. However, as the external temperature pushes the notches / peaks inside the laser line (or if it is located there upon start-up), and as the error signal becomes negative, the controller “reverses sign” and begins to lock on the alternate set point on the left of the notches / peaks where again the condition IpDmoni = IpDmon2 exists as shown in Figure 19B. A new thermal set point is reached as shown in the region between 0.010s and 0.015s. The only caveat to this is that the electrical to optical bit pattern is now inverted, such that an electrical 0 is an optical 1 (high) and an electrical 1 is an optical 0 (low). However, this should not significantly impact the data path as this can be easily reversed later in the data path.

[0159] The result of this effective absolute value (which may also be accomplished in other ways with offset biases) is to change the simple PID controller’s “dead-zone”

[0185] into a useful locking area as well. As shown in Figure 28A, the laser line

[0187] lies in the middle ofthe notches / peaks and since there is no active cooling the notches / peaks at start-up cannot be shifted towards to left. In this area, the error signal in the PID controller is non-zero, but negative. If the amplifier in Figure 26C is used, the new lock-point becomes the left-most set of curves where the currents are equal again

[0189] . A second circumstance of this absolute valuing is that the since the lock-points are on the complementary optical thru and drop powers, the optical bit pattern that is generated is the complementary bit patten and 0’s become l’s as l’s become 0’s. But very standard bit inversion circuits can be used to treat these signals.

[0160] Another method that can be used to make sure that the laser line does not coincide with the middle of the notches / peaks at start-up is to use a thermal kick at the start-up.

[0161] Figure 29 shows a simple circuit that can be used at the output driver of Figure 26B or 26C to allow for a momentary increase in thermal resistor current should the error signal indicate the laser line is in the dead zone. As the output goes towards zero, the diode becomes closed circuit and supplies small current to heater allowing it to pass the notch. Once this happens, the normal operation of the PID controller can take over and the notches / peaks will be driven to the next FSR group where they will lock on the points where the currents are again equal.

[0162] Specific details are given in the above description to provide a thorough understanding of the embodiments. However, it is understood that the embodiments may be practiced without these specific details. For example, circuits may be shown in block diagrams in order not to obscure the embodiments in unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail in order to avoid obscuring the embodiments.

[0163] The foregoing disclosure of the exemplary embodiments of the present invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many variations and modifications of the embodiments described herein will be apparent to one of ordinary skill in the art in light of the above disclosure. The scope of the invention is to be defined only by the claims appended hereto, and by their equivalents.Further, in describing representative embodiments of the present invention, the specification may have presented the method and / or process of the present invention as a particular sequence of steps. However, to the extent that the method or process does not rely on the particular order of steps set forth herein, the method or process should not be limited to the particular sequence of steps described. As one of ordinary skill in the art would appreciate, other sequences of steps may be possible. Therefore, the particular order of the steps set forth in the specification shouldnot be construed as limitations on the claims. In addition, the claims directed to the method and / or process of the present invention should not be limited to the performance of their steps in the order written, and one skilled in the art can readily appreciate that the sequences may be varied and still remain within the spirit and scope of the present invention.

Claims

CLAIMSWhat is claimed is:

1. A circuit comprising: an optical micro-ring resonator operating over a range of wavelengths; a plurality of photodetectors optically coupled to the optical micro-ring resonator; and a closed-loop control circuit receiving electrical signals generated by the plurality of photodetectors; wherein the closed-loop control circuit tracks and locks at least one of a notch within a transmission response of the optical micro-ring resonator and a peak within the transmission response of the optical micro-ring resonator to a wavelength of an input optical signal coupled to the optical micro-ring resonator.

2. The circuit according to claim 1, wherein the plurality of photodetectors are a pair of photodetectors; and the electrical signals generated by the plurality of photodetectors represent a differential optical signal between an output port of the optical micro-ring resonator and another output port of the optical micro-ring resonator.

3. The circuit according to claim 1, wherein the plurality of photodetectors are a pair of photodetectors; and the electrical signals generated by the plurality of photodetectors represent a differential optical signal between an output port of the optical micro-ring resonator to which the input optical signal is being coupled in order to drop the input optical signal and another output port of the optical micro-ring resonator representing a through port comprising other optical signals not being dropped by the optical micro-ring resonator.

4. The circuit according to claim 1 , wherein the closed-loop control circuit employs an auto-tuning servo control loop to continuous lock the optical micro-ring resonator to the wavelength of the input optical signal.

5. The circuit according to claim 4, wherein the closed-loop control circuit local heat a portion of the optical micro-ring resonator circuit to establish a minimum value of an error signal.

6. The circuit according to claim 1, wherein the optical micro-ring resonator incorporates a p-n junction to enable modulation of an output of the optical micro-ring resonator in dependence upon an applied electrical modulation signal; and the closed-loop control circuit employs an auto-tuning servo control loop to continuous lock the optical micro-ring resonator to the wavelength of the input optical signal by locally heating a portion of the optical micro-ring resonator circuit to establish a minimum value of an error signal.

7. The circuit according to claim 1, wherein the closed-loop control system locks to the at least one of the notch within a transmission response of the optical micro-ring resonator and the peak within the transmission response of the optical micro-ring resonator independent of any measurement of the wavelength of the input optical signal or data being provided to the closed-loop control system indicative of the wavelength of the input optical signal.

8. A circuit comprising: a closed-loop control system receiving complementary signals generated in dependence upon signals from photodetectors coupled to an optical micro-ring resonator; wherein the closed-loop control system locks a feature of a transmission characteristic the optical microring resonator a wavelength of an input optical signal independent of any measurement of the wavelength of the input optical signal or data being provided to the closed-loop control system indicative of the wavelength of the input optical signal.

9. The circuit according to claim 8, wherein the feature of the transmission characteristic is a mid-point in the transmission characteristic; the complementary signals generated by the photodetectors relate to monitoring ports on outputs of the optical micro-ring resonator; and the closed-loop control circuit maintains the mid-point such that the optical signals within the outputs are balanced under electrical modulation of the optical micro-ring resonator.

10. The circuit according to claim 8, wherein the complementary signals generated in dependence upon signals from photodetectors are generated with processing circuits that each receive an electrical output of a photodetector and another electrical signal established by averaging an electrical modulation signal applied to the optical micro-ring resonator.

11. The circuit according to claim 8, wherein the complementary signals generated in dependence upon signals from photodetectors are generated with processing circuits that each receive an electrical output of a photodetector and another electrical signal established by averaging an electrical modulation signal applied to the optical micro-ring resonator; and each processing circuit incorporates a variable gain transimpedance amplifier such that the closed-loop control system can establish the lock independent of whether the electrical modulation signal is balanced or unbalanced.

12. The circuit according to claim 8, wherein the complementary signals generated in dependence upon signals from photodetectors are generated with processing circuits that each receive an electrical output of a photodetector and another electrical signal established by averaging an electrical modulation signal applied to the optical micro-ring resonator; each processing circuit incorporates a variable gain transimpedance amplifier such that the closed-loop control system can establish the lock independent of whether the electrical modulation signal is balanced or unbalanced; and the processing circuits support variable weighting functions when the optical micro-ring resonator forms part of a photonic matrix multiplier.

Images