Abstract
A system and method to convert a wideband optical signal to a multi-bit digital electrical signal using a superconducting integrated circuit. In a preferred embodiment, the optical signal modulates the phase (i.e., adjusts the timing) of a sequence of single-flux-quantum voltage pulses. The optoelectronic modulator may comprise an optically tunable Josephson junction, superconducting inductor, or bolometric detector, with switching speeds approaching 100 ps or less. The optical signal may comprise a plurality of optical signals such as a wavelength-division multiplexed signal. The optical-to-digital converter may be applied to high-speed digital communication switches, broadband digital input/output for superconducting or quantum computing, and control/readout of detector arrays.
Claims
1. A transducer for receiving an optical signal and generating a digital representation of information modulated on the optical signal having a modulation bandwidth, comprising: at least one optically-responsive superconductor circuit comprising a Josephson junction, configured to produce an pulsatile electronic signal corresponding to a modulation of the optical signal and having a pulse rate in excess of the modulation bandwidth; and a digital circuit configured to convert the pulsatile electronic signal into a multibit digital representation of the information.
2. The transducer according to claim 1, wherein the optical signal is generated based on a quantum computing operation of a qubit.
3. The transducer according to claim 1, wherein the digital circuit is further configured to use the multibit digital representation of the information to control a qubit.
4. The transducer according to claim 1, wherein the digital circuit comprises a plurality of Josephson junctions, configured to perform digital logic functions based on single-flux-quantum (SFQ) pulses.
5. The transducer according to claim 4, wherein the plurality of Josephson junctions are configured as rapid single flux quantum logic.
6. The transducer according to claim 1, wherein the information modulated on the optic al signal has a bandwidth of at least 1 GHz.
7. The transducer according to claim 1, further comprising a plurality of qubits operating at a first cryogenic temperature, and a superconducting classical control chip operating at a second temperature, difference from the first cryogenic temperature.
8. The transducer according to claim 1, wherein the pulse rate is linearly responsive to changes in intensity of the optical signal.
9. The transducer according to claim 1, wherein the optical signal is modulated with multiplexed information streams, further comprising an optical demultiplexer configured to demultiplex the multiplexed information streams of the optical signal.
10. The transducer converter according to claim 9, wherein the optical demultiplexer comprises a plurality of microring resonant waveguides.
11. The transducer according to claim 9, wherein the multiplexed information streams are wavelength division multiplexed on the optical signal.
12. The transducer according to claim 11, further comprising a respective microring resonant waveguide for each respective modulated wavelength of the wavelength division multiplexed information streams on the optical signal, wherein each respective microring resonant waveguide has a respective optically-responsive superconductor circuit, each respective optically-responsive superconductor circuit comprising a Josephson junction configured to produce an pulsatile electronic signal corresponding to a modulation of the optical signal with a respective multiplexed information stream and a digital circuit configured to convert the respective pulsatile electronic signal into a multibit digital representation of the respective multiplexed information.
13. The transducer according to claim 1, wherein the optical signal is modulated with multiplexed information streams comprising a plurality of respective information streams from a plurality of respective qubits, multiplexed at different respective frequencies.
14. The transducer according to claim 1, wherein the optically-responsive superconductor comprises a kinetic inductance bolometer.
15. A method for transducing an optical signal to a digital representation of information modulated on the optical signal having a modulation bandwidth, comprising: producing an pulsatile electronic signal corresponding to a modulation of the optical signal and having a pulse rate in excess of the modulation bandwidth with at least one optically-responsive superconductor circuit comprising a Josephson junction; and converting the pulsatile electronic signal into a multibit digital representation of the information with a digital circuit.
16. The method according to claim 16, wherein the optical signal is generated based on a quantum computing operation of a qubit.
17. The method according to claim 16, further comprising controlling a qubit with the multibit digital representation of the information.
18. The method according to claim 16, wherein the pulse rate is linearly responsive to changes in intensity of the optical signal.
19. The method according to claim 16, further comprising demultiplexing a plurality of frequency-multiplexed information streams of the optical signal with a plurality of microring resonant waveguides, wherein each respective microring resonant waveguide has a respective optically-responsive superconductor circuit, each respective optically-responsive superconductor circuit comprising a Josephson junction configured to produce an pulsatile electronic signal corresponding to a modulation of the optical signal with a respective multiplexed information stream and a digital circuit configured to convert the respective pulsatile electronic signal into a multibit digital representation of the respective multiplexed information.
20. A transducer for receiving a wavelength division multiplexed optical signal and generating digital representation of information modulated on different channels of the wavelength division multiplexed optical signal, comprising: a plurality of microring resonators; and a plurality of optically-responsive superconductor circuits, each respective superconductor circuit comprising: a Josephson junction, configured to receive an output of a respective microring resonator and, to produce an pulsatile electronic signal having a pulse rate in excess of a modulation bandwidth of a respective channel of the wavelength division multiplexed optical signal and corresponding to the information modulated on the respective channel of the wavelength division multiplexed optical signal; and a digital circuit configured to convert the pulsatile electronic signal into a multibit digital representation of the information modulated on the respective channel of the wavelength division multiplexed optical signal.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0083] FIGS. 1A and 1B show the characteristics of a Josephson junction of the prior art. FIG. 1A shows the circuit symbol together with the time-average voltage <V>, and FIG. 1B shows fast generation of voltage pulses V(t).
[0084] FIG. 2 shows a Josephson junction generator and pulse transmission line of the prior art.
[0085] FIG. 3 shows a block diagram of a prior-art superconducting phase-modulation-demodulation analog-to-digital converter (ADC).
[0086] FIG. 4 shows a conceptual diagram of prior-art micro-ring optical resonators for a multiplexer and demultiplexer of a wavelength-division multiplexed (WDM) optical signal.
[0087] FIG. 5 shows a block diagram of a preferred embodiment of an optical-to-digital converter.
[0088] FIGS. 6A and 6B shows block diagrams of two alternative embodiments of an optically modulated voltage pulse generator: a direct optical pulse-rate modulator (FIG. 6A), and an optically modulated pulse delay (FIG. 6B).
[0089] FIG. 7 shows examples of dependence of voltage pulse rate on optical intensity.
[0090] FIGS. 8A and 8B show a symbolic representation of a variable Josephson junction (FIG. 8A), and an I-V plot (FIG. 8B), where the critical current I.sub.c is dependent on optical intensity.
[0091] FIG. 9 shows a Josephson pulse generation/propagation circuit, where one or more Josephson junctions and/or inductors may be optically varied.
[0092] FIGS. 10A and 10B show alternative embodiments of a variable Josephson pulse generator.
[0093] FIG. 11 shows a circuit diagram of one embodiment of Josephson delay line and synchronizer circuit for a superconducting optical-to-digital converter.
[0094] FIG. 12 shows a conceptual diagram of an integrated optical demultiplexer and array of superconducting optical-to-digital converters for a WDM optical signal.
[0095] FIG. 13 shows a block diagram for a clock recovery and word bit alignment circuit that may be used with an array of optical-to-digital converters such as that in FIG. 12.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0096] FIGS. 1-4 describe prior-art technology of superconducting and optical devices that may be used in preferred embodiments of the proposed new optical-to-digital converter.
[0097] FIGS. 1A and 1B describe the electrical properties of a Josephson junction of the prior art, indicated in a circuit by an “X”. A Josephson junction is a weak superconductor link connecting two superconducting wires. For a current below a characteristic critical current L (which may be of order 0.1-1 mA), a Josephson junction acts as an ideal lossless superconductor. For larger currents, the Josephson junction develops a voltage and becomes resistive, with a typical resistance of order 1-10 ohms, as shown in the dc I-V curve of FIG. 1B.
[0098] Strictly speaking, the Josephson junction is a fast dynamic device, and the plot of FIG. 1B shows the instantaneous voltage V(t), which gives rise to the average voltage <V> in FIG. 1A. For current just above the critical current, V(t) corresponds to a series of fast voltage pulses, each pulse having integrated voltage of one flux quantum Φ.sub.0=h/2e=2.07 mV-ps. An actual SFQ (single-flux-quantum) pulse is rounded rather than rectangular (as in FIG. 1B) with a typical pulse height of order 1 mV, and effective pulse width of order 2 ps, depending on circuit parameters. The pulse frequency is f=<V>/Φ.sub.0.
[0099] A Josephson junction is also characterized by an effective capacitance C and nonlinear inductance L.sub.J, where typically L.sub.J˜10 pH and C˜1 pF. This would correspond to an LC resonator, except that it is damped by the resistance R. An underdamped Josephson junction tends to ring, and has an undesirable hysteretic I-V curve. An overdamped Josephson junction tends to have wide pulses, slowing down the dynamic operation. The optimum mode of operation is near critical damping, where the Q of the resonator is near 1. A physical Josephson junction may be based on a niobium tunnel junction, and tends to be underdamped. A practical Josephson junction may comprise a tunnel junction with a parallel resistive shunt added to bring the combined device near critical damping. The Josephson junctions discussed herein should be assumed to be damped junctions, which may include a resistive shunt. Other types of damping may be used.
[0100] FIG. 2 shows the operation of a Josephson transmission line of the prior art (see, for example, www.physics.sunysb.edu/Physics/RSFQ/Lib/PB/jtl.html), whereby a single SFQ pulse generated by the first Josephson junction on the left is regenerated by each junction in the sequence, thus propagating the pulse down the line. The generating junction is biased just above its critical current, while the propagating junctions are biased somewhat below their critical current (typically at ˜70%). In addition to the Josephson junctions, the JTL comprises connecting inductances, where this inductance is a superconducting wire or microstripline. The inductance per stage is typically L=Φ.sub.0/I.sub.c, since a larger inductance can cause an SFQ pulse to form a static trapped flux in the loop, rather than propagating. The propagation time through the JTL depends on both I.sub.c and L, and is typically several ps per stage. The JTL is a primary element used to connect gates in rapid-single-flux-quantum (RSFQ) logic, which may be applied to a preferred embodiment of the present invention.
[0101] FIG. 3 shows a detailed block diagram of a phase-modulation-demodulation (PMD) ADC of the prior art. The preferred embodiment of this technology is based on RSFQ logic, but this could alternatively be implemented using another device technology. A rapidly varying analog signal (bandwidth up to ˜1 GHz) enters on the left, and it modulates the pulse output of the “flux quantizer”. A positive signal could increase the pulse rate, while a negative signal could decrease the rate. The resulting modulated pulse train enters the “synchronizer” block, which acts to demodulate the pulse sequence to recover a digital representation of the time-derivative of the input analog signal, at a sampling rate given by the main clock. For example, the unmodulated pulse rate could be 10 GHz, and the main clock could be 20 GHz. The quantizer and synchronizer together comprise the PMD ADC modulator. This derivative is then integrated (or summed) to regenerate a digital representation of the analog signal, and averaged to increase the effective number of bits and decrease the noise level. The clock for the integration and averaging may be a factor of N smaller than the main clock rate. The integrator and averager together may comprise a digital decimation filter. In a preferred embodiment, the ADC modulator and at least a portion of the digital decimation filter may be integrated on the same chip. The optical-to-digital converter of the present invention may employ some similar circuits to those in the phase modulation-demodulation analog-to-digital converter (PMD ADC).
[0102] FIG. 4 presents a conceptual diagram of a wavelength-division multiplexed (WDM) optical communication system of the prior art. In this example, the transmitter comprises an optical carrier signal that is modulated by four separate electrical signals, preferably at slightly different wavelengths so as not to interfere. At the receiver, these four electrical signals are separated using a demultiplexer, and detected in parallel. The multiplexer and demultiplexer may comprise an array of integrated optical ring resonators, coupled to optoelectronic modulators.
[0103] FIG. 5 shows a block diagram of a preferred embodiment of the optical-to-digital converter of the invention. A voltage pulse generator is modulated by an optical source. The asynchronous pulse train is then synchronized, integrated, and averaged in much the same way as PMD ADC. The unmodulated pulse rate might be 10 GHz, and the sampling clock might be twice this or 20 GHz. The integrator and averager could be combined into a single digital filter, or split into stages that are done in different technologies. The averaging clock f.sub.av determines the effective bandwidth of the digitizer; by the Nyquist relation, BW=f.sub.av/2.
[0104] FIGS. 6A, 6B, and 7 address in general terms how a digital optical modulator might work. Either the pulse rate could be directly modulated by the light, or the pulse delay could be modulated by the light. FIG. 6A shows a mode whereby the pulse rate is directly modulated by the optical input. The pulse rate might either increase or decrease with increasing optical power, as shown in FIG. 7. In ideal cases, the change in the pulse rate AR may be linearly proportional to the optical power P. Linearity is preferred in communication systems, because nonlinearity tends to produce frequency mixing, generating frequency components where they would not otherwise be present. An alternative to direct optical modulation is shown in FIG. 6B, where the pulse delay associated with pulse propagation is either advanced or retarded by the optical power. In this case, a constant optical power would not change the average pulse rate, but a transient change in optical power dP/dt would cause a proportional change in pulse rate AR. In either case, the digital representation of the optical power P(t) could be recovered, but the appropriate digital processing would differ.
[0105] FIG. 8A shows a symbol for a variable Josephson junction (a line with an arrow passing through the “X”), whereby at least one of the junction parameters could change under optical illumination. In one embodiment shown in the set of I-V curves in FIG. 8B, the critical current I.sub.c of the junction is decreased under optical illumination. For a fixed bias current with I>I.sub.c, a decrease in I.sub.c would cause an increase in voltage, and hence an increase in the generator pulse rate. For a current bias with I<I.sub.c, a small optical power will change the current-dependent Josephson inductance, which can change the delay of a junction in pulse-propagating mode. Alternatively, in another embodiment, the junction resistance could change due to optical illumination, which could also change the pulse generation rate for I>I.sub.c. There may be several mechanisms that could cause optical modulation in a Josephson junction. The critical current is normally a function of temperature, so that any type of nonequilibrium heating would also cause I.sub.c to decrease. Alternatively, optical illumination might affect the tunnel barrier connecting the two superconducting electrodes, which could affect both I.sub.c and R.
[0106] FIG. 9 shows the prior-art Josephson transmission line, where now one or more elements can be modulated by optical illumination, which could modulate the pulse generation rate, and/or the pulse delay rate. The modulated elements could include a Josephson junction, as well as a connecting superconducting inductor. A ring oscillator configuration may also be implemented.
[0107] A superconducting inductor will include at least a portion of its inductance as “kinetic inductance” associated with kinetic energy stored in the superconducting charge carriers. The kinetic inductance is dependent on temperature, and therefore on any nonequilibrium heating that affects the superconducting electrons. A very thin superconducting film of thickness d<λ.sub.L (the magnetic penetration depth of the superconductor) exhibits an enhanced kinetic inductance. Specifically, the kinetic inductance of a thin film per square is μ.sub.0λ.sub.L.sup.2/d. So for example, if d=10 nm and λ.sub.L=100 nm, the kinetic inductance is of order 1 pH/square, and tends to dominate the usual magnetic inductance.
[0108] FIGS. 10A and 10B show how optical illumination may affect other superconducting circuit elements. FIG. 10A represents a SQUID, comprising two Josephson junctions in an inductive loop. The SQUID is equivalent to a single Josephson junction with a variable critical current. Even if the critical currents of the component junctions remain unchanged, the critical current of the SQUID is dependent on the flux contained in the superconducting loop. The flux can be varied by varying the loop inductance, which may be sensitive to optical illumination.
[0109] FIG. 10B represents a resistively shunted Josephson junction with an optically sensitive shunt resistance. This resistance might comprise a superconducting film with a critical temperature T.sub.c smaller than that of the Josephson junction. For example, the Josephson junction and connecting inductors might comprise niobium (Nb) with critical temperature T.sub.c=9.2 K, while the resistor might comprise tantalum (Ta) with T.sub.c=4.3 K. If the operating temperature is close to T.sub.c of the Ta film, its resistance will be sensitive to optical illumination that may produce transient heating. In these ways, these circuits can be used to create variable pulse generating or propagating elements.
[0110] FIG. 11 shows a detailed circuit diagram of one preferred embodiment of a PMD optical modulator using RSFQ logic, similar in part to the PMD modulator of FIG. 3. The master clock generates a periodic pulse stream, which is fed to one or more JTL sections. Each section provides a small time delay, and these time delays may be modulated by light incident on one or more of the elements, either the Josephson junctions or the inductors. This represents a delay-type modulator, as shown in FIG. 6B. The asynchronous delay-modulated pulse stream then enters the synchronizer, which is labeled here as a race-arbiter-type phase demodulator. This synchronizer, quite similar to those in prior-art PMD ADCs, uses a set of data-flip-flops (DFF) to compare the arrival of clock pulses and data pulses. The DFF is a standard RSFQ cell that is also known as an RS flip-flop (for reset-set). See, for example, www.physics.sunysb.edu/Physics/RSFQ/Lib/AR/dff.html. This circuit is a multichannel synchronizer that uses a set of calibrated time delays (based on JTLs) to subdivide the clock period into 4 quantization levels, thus increasing the timing precision and increasing the effective resolution of the ADC.
[0111] FIG. 12 shows a conceptual diagram of an integrated multi-channel optical-to-digital receiver. This would be relevant for receiving and digitizing a WDM optical signal, comprising multiple signals modulated at slightly different wavelengths. The figure shows a demultiplexer that permits each signal to be coupled out using a set of integrated microring resonant waveguides. The waveguides are resonant for each wavelength, indicated in the figure by different size rings. Each of the wavelengths is coupled to a superconducting optical-to-digital converter, and each of the digital signals is sent out in parallel. In a preferred embodiment, both the demultiplexer array and the ADC array could be integrated on the same chip.
[0112] In order to correlate the various signals at their proper times, FIG. 12 also shows a clock recovery and word bit alignment circuit, also known as word synchronization.
[0113] FIG. 13 shows one preferred embodiment of a word synchronization circuit, comprising a word transmit circuit on the chip, and a word receive circuit some distance away. Word synchronization circuits are known in other technologies, and are disclosed for RSFQ circuits in U.S. Pat. No. 9,520,180. In FIG. 13, D is the D-flip-flop, N is the inverter, C is the Muller C-element, and FIFO is a first-in-first-out shift register, all of which are standard RSFQ cells.
[0114] Such a superconducting optical-to-digital converter may be applied to general optical fiber communication systems, or alternatively to data transfer to other superconducting or cryogenic circuits. Such superconducting or cryogenic circuits may include RSFQ computing systems, cryogenic detector arrays, or cryogenic implementations of quantum computing. In many cases, signal bandwidths in excess of 1 GHz, or even in excess of 10 GHz, may be required. RSFQ digital processing is capable of these data rates, and properly implemented optoelectronic superconducting modulators also have sufficient bandwidth. For example, ultrathin superconducting films can change the effective electron temperature at GHz frequencies, without producing wider heating that would slow the response.
[0115] The embodiments presented here are not exclusive, but are used to illustrate the range of circuits and devices that may be functional components of superconducting optical-to-digital converters.
[0116] There has thus been shown and described systems and methods for optical-to-digital conversion which fulfill all the objects and advantages sought therefor. Many changes, modifications, variations, combinations, subcombinations and other uses and applications of the subject invention will, however, become apparent to those skilled in the art after considering this specification and the accompanying drawings which disclose the preferred embodiments thereof. All such changes, modifications, variations and other uses and applications which do not depart from the spirit and scope of the invention are deemed to be covered by the invention, which is to be limited only by the claims which follow.
