Photonic integrated circuits and methods

Abstract

To provide a photonic integrated circuit (PIC) used to generate a random number.SOLUTION: The PIC includes on a substrate: a light source 2; a first photodetector 4 for outputting an electrical signal used when generating a random number in response to receiving light from the light source 2; and a light guidance system 6 for directing light from the light source 2 to the photodetector 4.EFFECT: An genuine random number is generated.SELECTED DRAWING: Figure 1

Description

Background Art

[0001] Random number generators based on the inherently random behavior of photons caused by the effects of quantum mechanics are known. For example, the behavior of a single photon can be determined and used to generate true random numbers.

[0002] It is desirable to improve random number generators based on photon detection.

Brief Description of the Drawings

[0003] [Figure 1] FIG. schematically shows a system for generating random numbers based on an example. [Figure 2] FIGS. a and b show plots related to the current readings of a photodetector according to an example. [Figure 3] FIG. is a schematic cross-sectional view of a photonic integrated circuit (PIC) of an example (for example, the example of FIG. 1). [Figure 4] FIG. is a flowchart illustrating a method for generating random numbers based on an example. [Figure 5] FIG. illustrates a manufacturing method of an example.

Modes for Carrying Out the Invention

[0004] The examples described herein relate to photonic integrated circuits (PICs) used to generate random numbers based on the behavior of photons due to quantum mechanical effects. More specifically, a PIC can be considered a monolithic PIC when the optical components for generating random numbers are formed on a single substrate (in other words, on a common substrate). The substrate is, for example, a compound of elements from groups III and V of the periodic table, such as a so-called III-V semiconductor compound, such as indium phosphide (InP). Such a PIC can be completely housed without optical input or output and is compact enough to be integrated into devices such as computers and smartphones. In practice, such a PIC is more compact and simpler in configuration and operation than known random number generators. For example, a PIC can be 100 microns × 1000 to 4000 microns × 4000 microns or less.

[0005] Figure 1 schematically illustrates an example of a system for generating random numbers. This system is implemented, at least in part, as a PIC on a common substrate, for example, a III-V semiconductor compound (e.g., InP). Figure 4 illustrates a flowchart of the method for generating random numbers.

[0006] The PIC used to generate random numbers includes a light source 2 on a substrate (illustrated later using Figure 3), a photodetector 4 on the substrate, and an optical guidance system 6 on the substrate. The optical guidance system is configured to direct light from the light source to the photodetector. The waveguide 8 of the optical guidance system (illustrated schematically as an arrow) guides light from the light source to the photodetector 4. When light from the light source enters the photodetector 4, in other words, when it is received by the photodetector 4, the photodetector outputs an electrical signal used to generate random numbers. This will be explained in more detail later. The light source is, for example, a semiconductor optical amplifier (SOA) or a laser. The light detector, in other words, a photodetector, is, for example, a photodiode capable of detecting low light intensity. Such a photon detector is simpler than known examples called "single-photon detectors," which use more complex or design-limit-bound detectors, such as silicon photomultiplier tube (SiPM) avalanche photodiodes. Furthermore, the photodiode in the example is a monolithic integrated photodiode with a sampling rate of at least gigahertz (GHz) frequency. This is the same as the light in known systems. This enables faster operation than detectors, and therefore faster random number generation. Furthermore, such photodiodes with GHz or higher sampling rates can detect photocurrents larger than those of known photodetectors. As a result, background electron noise is further overcome, and fluctuations related to quantum effects become even more dominant.

[0007] In examples such as those shown in Figures 1-3, an optical attenuation system is located on a common substrate. The attenuation system is configured to reduce the intensity of the light output from the light source to the intensity required by the photodetector 4 used to generate random numbers. Thus, the attenuation system reduces the light intensity between the light source and the photodetector 4. In examples such as those shown in Figures 1-3, the attenuation system includes one or more optical splitters 10 (e.g., passive optical splitters) on the common substrate. Figure 1 shows two such optical splitters, but this is schematic. The first optical splitter 10a receives light from the light source before the other optical splitter (e.g., the second optical splitter 10b) receives light from the light source (through the first optical splitter). Figure 1 is schematic and shows the first and second optical splitters, but in some other examples there may be only one such optical splitter, and in other examples there may be more than two optical splitters as needed to reduce the light intensity to a level suitable for the photodetector 4 (and the overall system setup) to be used when generating random numbers.

[0008] A suitable level is low enough that any fluctuation in light intensity detected by the photodetector 4 is governed by quantum effects and therefore follows Poisson statistics. In other cases, if the intensity is higher, and therefore a larger number of photons per sample are detected by the photodetector, any fluctuations resulting from quantum effects are masked and therefore negligible compared to the total number of photons detected per sample, and thus cannot be sufficiently detected and used for random number generation. Those skilled in the art are well familiar with Poisson statistics and, by extension, the Poisson probability distribution. The Poisson probability distribution describes the probabilities of several events occurring during a given time. The events occur at a known constant average rate and are independent of the time since the last event. The application of the Poisson distribution in random number generation will be discussed further below.

[0009] Typically, the standard deviation (or intrinsic variability) of a Poisson distribution is the square root of the mean sample size. Therefore, the variability with respect to the mean sample size decreases as the number of photons per sample increases. On the other hand, at low photon counts (especially single-photon counting), optimized photodiodes and electronics are required. InP's PIC features a high sample rate and low electronic background noise in the photodiode. Therefore, at a photon count of 1000 photons per sample, the intrinsic quantum fluctuations still far exceed the electronic background current (at least about 10 times). At high sample rates, such 1000 photons per sample still generate a considerable photodiode current that can be measured using conventional readout electronics.

[0010] The light source, photodetector, and optical splitter of the optical attenuation system are optically connected to each other by an optical guidance system (e.g., waveguide 8) as needed. Thus, the input of the first optical splitter is optically connected by a waveguide to receive light from the light source (directly, in some examples), and the output of the first optical splitter is optically connected by a waveguide to either the input of a subsequent optical splitter or a photodetector. Each splitter can be thought of as guiding or redirecting a portion of the light from its input to its output, but since the function of a splitter is to reduce the intensity of light propagating from the light source, the splitter is described herein as part of the optical attenuation system.

[0011] In some examples, the light attenuation system works by attenuating light from the light source between the light source and the photodetector 4. It is configured to attenuate the intensity to at least 1 / 1000th of its original level. Therefore, for example, the intensity of light incident on the photodetector 4 is at most 0.1% of the intensity output by the light source.

[0012] As those skilled in the art will see, such attenuation can be achieved in different ways. The example described herein uses optical splitters. Each optical splitter splits the light generated from a light source into a first portion of light to be propagated by an optical guidance system to the photodetector 4, and a second portion of light that is not guided to the photodetector 4. Multiple optical splitters may be arranged in series with one another. The output of the first portion of light from one optical splitter is guided forward and input to the next optical splitter in series, which then outputs its first portion of light and guides it forward to any subsequent optical splitter in series. This continues until it reaches the photodetector 4.

[0013] Each optical splitter in the attenuation system is, for example, an optical filter, or, for example, a multimode interferometer (MMI), which is well known to those skilled in the art. In an MMI, for example, the waveguide width is widened to transition the light from single-mode propagation to multimode propagation. The spatial distribution of optical modes changes as a function of the MMI length, and the appropriate distribution required for a particular splitter application can be calculated using numerical calculations. In the case of a 1×2 splitter, two waveguides are fabricated for a length from the start of the MMI (where the two modes are formed and spatially sufficiently separated).

[0014] Depending on the specifications of each splitter, in some cases each splitter reduces the light intensity by 50% (the first portion of the light input to the splitter is 50% intensity and the second portion is 50% intensity), and in other cases each splitter reduces the light intensity by 70% (the first portion of the light input to the splitter is 30% intensity and the second portion is 70% intensity). In the former case, in some cases ten so-called 50:50 splitters are used in series to reduce the intensity to the photodetector 4. In the latter case, in other cases six so-called 30:70 splitters are used in series to reduce the intensity to the photodetector 4. Further examples may include different splitting ratios.

[0015] The MMI is an example of a passive optical splitter. A passive optical splitter does not require voltage drive to function (as in the case of an active component). Using a passive optical splitter in an optical attenuation system means that higher power can be used for the light source, and as a result, the need for and / or power requirements of the circuit to drive active components elsewhere in the PIC can be reduced. Furthermore, as will be discussed further below, such a passive optical splitter means that the intensity of the light output by the light source can be monitored, and as a result, it can help calibrate and adjust the system for generating random numbers.

[0016] The photodetector 4 described herein is also referred to as the first photodetector. In addition, in some examples, such as the example in Figure 1, there are one or more second photodetectors 12, each optically connected by a waveguide 8 to receive the second portion of the optical output from the corresponding optical splitter 10. Each second photodetector is a photodiode of the same type as the first photodetector, although in other examples they may be of a different type. For example, the first photodetector may be a balanced photodiode or another type of optimized high-precision photodiode, and the second photodetectors may be the same or a different type of photodetector.

[0017] The PIC includes a first circuit for processing electrical signals output from each of one or more second photodetectors. Such second photodetectors are used, for example, to monitor the intensity of light emitted by a light source. If a second photodetector detects a deviation from the target intensity, this indicates that the intensity of the light in the first portion guided forward to the photodetector is also deviating, allowing appropriate action to be taken. For example, one or more of the second photodetectors For each of these, the first circuit receives an electrical signal from the corresponding second photodetector. The characteristics of the electrical signal (e.g., current) are measured to determine the magnitude of a characteristic that indicates the actual intensity of the light output by the light source. If it is determined that the actual intensity of the light output by the light source differs from the target intensity of the light that the light source should output (for example, based on the requirement that the second portion of light must be 50% of the intensity of the light output by the light source), the first circuit can take appropriate action to maintain the ability of the random number generation system. Such a response may involve adjusting the current applied to the light source to adjust the intensity of the light output by the light source. Thus, if the intensity of the second portion is less than the target intensity, the current can be increased to increase the intensity of the light output by the light source and reduce the difference between the actual intensity and the target intensity. Furthermore, or alternatively, such a response may involve calibrating the random number generation process (described further below) (by adjusting a threshold or at least one of the characteristics measured from the electrical signal output by the first photodetector 4 (e.g., current)) to compensate for any deviation in the light intensity from the light source. Otherwise, such deviations could cause the system to malfunction, and any estimated random numbers generated might not actually be truly random. Such monitoring is performed at least by a second photodetector that receives the second portion of light from a first optical splitter in series, and in some examples, attenuation is monitored using multiple portions of the second portion of light from the optical splitters in series. For example, the second portion of light from each optical splitter is output to different second photodetectors, respectively, to monitor each attenuation step by the splitter, and any deviation from the target intensity is adjusted in the processing of the current applied to the light source and / or the electrical signal output by the first photodetector 4.

[0018] Next, we will describe the generation of random numbers. The PIC includes a second circuit configured to use the electrical signal output by a first photodetector 4 for generating random numbers. Generally, the second circuit receives the electrical signal output by the photodetector 4 and determines the value to use when generating random numbers based on the characteristics of the electrical signal (e.g., current). In these examples, such values ​​are binary values ​​(otherwise known as bits) and consist of values ​​of 1 or 0. A string of binary values ​​may also be generated based on a series of measurements of current (one binary value per measurement). The characteristics of the measured electrical signal change (or fluctuate) over time. This is caused by quantum effects that affect the passage of photons from the light source to the first photodetector 4. This means that the number of photons incident on the photodetector 4 is different at different points in time, and this results in fluctuations in the characteristics of the output electrical signal. Such fluctuations are truly random because they are determined by quantum effects and can be used as the basis for determining each binary value, thereby generating a truly random string of binary values. Such a string can then be used to generate a truly random number, for example, a decimal integer.

[0019] Various methods can be considered to generate random numbers by utilizing the variations in the characteristics of the electrical signal output by the first photodetector 4. For example, since the variations are determined by quantum effects, the measured values ​​fit a Poisson distribution. Therefore, random numbers can be generated by comparing each measured value of the electrical signal characteristics to a Poisson distribution or the parameters derived therefrom, and determining the value obtained for each measured value (e.g., 1 or 0).

[0020] Figure 2a shows an example plot of many current measurements of the electrical signal output by the photodetector 4. The plot assumes that 1000 photons are detected per measurement (referred to as readout in the figure). The x-axis represents time, and the y-axis represents the magnitude of the current for each readout. Each readout can be considered a sample of current, and subsequent samples are taken periodically at equal duration intervals that are separated from each other. As can be seen from Figure 2a, the current per readout fluctuates around an average magnitude, illustrated by the concentration of data points along the horizontal band on the plot. Figure 2b shows the same readouts, here plotted as a histogram. The x-axis shows the magnitude of the current for each readout, and the y-axis shows the frequency of each measurement. The histogram plot follows a Poisson distribution. Determine the mean readout (in these examples, the mean current magnitude). This mean is the average value that corresponds to the most commonly measured current magnitude, for example, as shown by the vertical line 14 in Figure 2b. Such an mean can be set as a threshold to determine which value to assign for use in generating random numbers, or, in other examples, a different statistical value can be calculated and used to set the threshold. After determining the mean, associate each subsequent measurement of a characteristic such as current with the mean. For example, calculate the difference between the measurement and the threshold, or determine whether the measurement is greater than or less than the threshold. The magnitude of the difference, or simply whether the measurement is greater than or less than the threshold, can be used to determine the value to assign from the measurement and generate random numbers. For example, Figure 2b shows that if the measurement is less than the threshold 14, a bit value of 0 is determined and recorded in the bit string, while if the measurement is greater than the threshold 14, a bit value of 1 is determined and recorded in the bit string. This method is repeated for each subsequent measurement to generate a string of randomly generated bit values, given that the variation in light intensity on the photodetector 4 depends on quantum effects. The interval separating each subsequent measurement from the previous measurement (and therefore the sampling frequency) may be set by the specific photodetector used, or it may be controlled by a second circuit based on a clock signal. Thus, the second circuit in the example receives an electrical signal output by the first photodetector 4, measures the characteristics of the electrical signal (e.g., current) at a first time point, and, based on the measured characteristics, determines if the magnitude of the measured characteristics is less than a threshold, and based on this, determines a first value (e.g., a bit value, e.g., 0 in Figure 2b) to be used when generating random numbers. At a second time point immediately following the first time point but separated by the sampling period time, a similar process is performed for the next measurement. However, instead, if the second measurement of the characteristics is greater than the threshold, the second value (e.g., a bit value, e.g., 1 in Figure 2b) is determined to be a different value from the first value.

[0021] In some examples, a threshold is determined before the circuit begins generating random numbers. Measurements of the characteristics of the electrical signal output by the photodetector 4 may be sampled over a given time long enough to determine the mean (e.g., the average) of the Poisson distribution of the measurements. This mean is then set as the threshold used when determining values ​​(e.g., bits) based on future variations in the characteristics. In other examples, the threshold may be determined and stored during a calibration process in manufacturing before shipping to consumers. This is again done by measuring the characteristics over a time long enough to determine the mean for the Poisson distribution.

[0022] As mentioned above, the intensity of the light output by the light source, and / or the light of any second portion output by the optical splitter 10, may deviate from the target intensity. Such deviations may be larger than any fluctuations caused by quantum effects. By monitoring any such deviations using one or more second photodetectors 12, the process for determining the values ​​for random number generation (e.g., bit values) may be adjusted accordingly. For example, an offset corresponding to the deviation may be added to the magnitude of the measured characteristic to compensate for the deviation and ensure a correct comparison with the threshold. Alternatively, the threshold may be adjusted to correspond to any such deviations to ensure a correct comparison of the measured magnitudes with respect to the threshold. Or, if any such deviations are sufficiently large or persist for a long period, the threshold may be reset by repeating the calibration process to determine the mean value according to the Poisson distribution.

[0023] In a specific example, we assume that the light source outputs light with an intensity of 200 microwatts (μW), the attenuation rate of the attenuation system is 1000, the sampling rate is 1 GHz, the threshold is 0.16 microamperes (μA), and the conversion efficiency of the first photodetector 4 in converting photons to electrons (and therefore measurable current) is 100%. Figure 3 shows an example of PIC layers. Cross-sectional views taken along lines A-A, B-B, C-C, and D-D in FIG. 1 are shown from left to right. A light source, waveguide 8, and photodiode 4 are each shown between an N-InP compound substrate and a p-type compound (e.g., P-InP, P-InGaAs, P-InAlAs, P-InAlGaAs, or P-InGaAsP (Ga is gallium, As is arsenic, Al is aluminum, and P is phosphorus)). The semiconductor materials based on the photodetector and the photodiode are the same in some examples but different in other examples. Any materials (e.g., electron conductivity or passivation to the environment) that are secondary to the functions described herein are not illustrated for clarity.

[0024] Based on this, the light intensity measured by the first photodetector 4 can vary by up to 3.1%. Since these variations are based on quantum effects, the range of readout magnitude values (of the current) covered by the Poisson distribution is 3.1%. Such a sufficiently large range of variations is useful for determining whether each measured value is greater than or less than a threshold value. This determination can become difficult as the range of variations decreases. The 3.1% range of variations is particularly larger than the range of variations obtained by systematic variations in a PIC without the optical attenuation system described herein. Note that the range of variations can increase due to a decrease in the intensity of the light output by the light source, a decrease in the conversion efficiency, an increase in the attenuation rate, and / or an increase in the sampling rate. In contrast, the range of variations can decrease due to an increase in the intensity of the light output by the light source, an increase in the conversion efficiency, a decrease in the attenuation rate, and / or a decrease in the sampling rate.

[0025] An electronic circuit (e.g., the first and second circuits described herein) may be implemented using at least one processor and at least one memory storing instructions. When the instructions are executed on at least one processor, they perform any of the example methods described herein in relation to at least the first or second circuit. The at least one processor may be, for example, a general-purpose processor, a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic circuit, discrete gates or transistor logic, discrete hardware components, or any suitable combination thereof, designed to perform the functions described herein. The first and second circuits may be provided together as a single printed circuit board (PCB), or each of the first and second circuits may be implemented on one or more substrates and connected to the PIC accordingly. As will be appreciated by those skilled in the art, suitable power circuits for powering the PIC and the first and second circuits are also contemplated.

[0026] As will be appreciated by those skilled in the art, various techniques may be used to deposit and pattern layers of semiconductor material according to the examples described herein. Such techniques may include chemical vapor deposition methods (e.g., metalorganic vapor phase epitaxy (MOVPE) or molecular beam epitaxy (MBE)). As will be appreciated by those skilled in the art, as part of the patterning, etching techniques may be used to remove portions of the material.

[0027] Accordingly, based on the example and with reference to Figures 501-504, the example PIC described herein is manufactured by a method comprising: preparing a substrate; forming a light source on the substrate; forming a first photodetector on the substrate, the first photodetector being configured to output an electrical signal used to generate random numbers in response to the reception of light from the light source; and forming an optical guidance system on the substrate configured to direct light from the light source to the first photodetector. Furthermore, the method in the example comprises forming an optical attenuation system on the substrate, the optical attenuation system being configured to attenuate the intensity of light from the light source between the light source and the photodetector so that the variation in the intensity of light detected by the photodetector is governed by Poisson statistics. Further in the example, the method comprises forming a plurality of the aforementioned second photodetectors, the optical attenuation system being a series of passive The system includes optical splitters, each of which is configured to split light from a light source into a first portion of light for propagation to a first photodetector by an optical guidance system, and a second portion of light for propagation to a second photodetector among a plurality of second photodetectors, which is used to adjust the intensity of light output by the light source by adjusting the current applied to the light source, thereby reducing the difference between the actual intensity output by the light source and the target intensity that the light source should output, or to calibrate a process for generating random numbers based on the difference between the actual intensity and the target intensity.

[0028] The examples described herein should be understood as illustrative examples only. Naturally, any feature described in relation to any one example may be used alone, in combination with other features described, or in combination with one or more features of any other example (or any combination of any other examples). Furthermore, equivalents and modifications not described herein may be used without departing from the scope of the appended claims.

Claims

1. A system used to generate random numbers, comprising a photonic integrated circuit, The aforementioned photonic integrated circuit is i) A light source on the substrate, ii) A photodetector on the substrate, configured to output an electrical signal used to generate random numbers in response to the reception of light from the light source, iii) An optical guidance system on the substrate, wherein the optical guidance system is configured to direct light from the light source toward the photodetector, iv) A light attenuation system on the substrate, The light attenuation system is configured between the light source and the photodetector to attenuate the intensity of light from the light source so that the fluctuation in the intensity of the light detected by the photodetector is governed by Poisson statistics. Includes, The aforementioned photonic integrated circuit a) Receiving the electrical signal output by the photodetector in ii), b) Measuring the characteristics of the electrical signal received in a) at a certain point in time, c) Based on the characteristics measured in b), it is determined that the magnitude of the characteristics measured in b) is smaller than the threshold, Based on d) and c), determine the value to be used when generating the random number, A system configured to perform the following actions.

2. The aforementioned substrate is A group III-V semiconductor compound, or at least one of indium phosphide or InP. The system according to claim 1.

3. The system according to claim 1, wherein the light attenuation system is configured to attenuate the intensity of light from the light source to at least 1 / 1000 between the light source and the photodetector.

4. The value used when generating the aforementioned random number is a binary value. The system according to claim 1.

5. The aforementioned time point is the first time point, the value used when generating the random number is the first value, and the photonic integrated circuit is e) Measuring the characteristics of the electrical signal received in a) at a second time point, Based on the characteristics measured in f) and e), it is determined that the magnitude of the characteristics measured in e) is greater than the threshold, Based on g) and f), determine a second value different from the first value used when generating the random number, It is configured to do, The system according to claim 1.

6. The second value is a binary value. The system according to claim 5.

7. The aforementioned characteristic is current. The system according to claim 1.

8. The threshold is the average magnitude of the characteristics of the electrical signal. The system according to claim 1.

9. a) Outputting light from a light source on the substrate, b) The light attenuation system on the substrate attenuates the intensity of light from the light source between the light source and the photodetector so that the fluctuation in the intensity of the light detected by the photodetector is governed by Poisson statistics, c) The light from the light source is guided to the photodetector on the substrate by the optical guidance system on the substrate, d) The photodetector outputs an electrical signal used to generate random numbers, e) Measuring the characteristics of the electrical signal received at a certain point in time, Based on the characteristics measured in f) and e), it is determined that the magnitude of the characteristics measured in e) is smaller than the threshold, Based on g) and f), determine the value to be used when generating the random number, including, A method for generating random numbers.

10. The aforementioned substrate is A group III-V compound, or at least one of indium phosphide or InP. The method according to claim 9.

11. The light attenuation system is configured to attenuate the intensity of light from the light source to at least 1 / 1000 between the light source and the photodetector. The method according to claim 9.

12. Determining the magnitude of an electrical signal output by a second photodetector, wherein the magnitude represents the actual intensity of the light output by the light source, Determining that the actual intensity of the light output by the light source is different from the target intensity of the light that the light source should output, At least one of the following: adjusting the current applied to the light source to adjust the intensity of the light output by the light source and reducing the difference between the actual intensity and the target intensity, or calibrating the process for generating the random numbers based on the difference between the actual intensity and the target intensity. including, The method according to claim 9.

13. The value used when generating the aforementioned random number is a binary value. The method according to claim 9.

14. The aforementioned time point is the first time point, the value used when generating the random number is the first value, and the method is, h) Measuring the characteristics of the electrical signal received at the second time point, i) Based on the characteristics measured in h), it is determined that the magnitude of the characteristics measured in h) is greater than the threshold, Based on j)i), determine a second value different from the first value to be used when generating the random number, including, The method according to claim 9.

15. The second value is a binary value. The method according to claim 14.

16. The aforementioned characteristic is current. The method according to claim 9.

17. The threshold is the average magnitude of the characteristics of the electrical signal. The method according to claim 9.

Images