Scalable parallel physical random number generator based on a superluminescent LED
Open accessAbstract
We describe an optoelectronic system for simultaneously generating parallel, independent streams of random bits using spectrally separated noise signals obtained from a single optical source. Using a pair of nonoverlapping spectral filters and a fiber-coupled superluminescent LED (SLED), we produced two independent 10 Gb/s random bit streams, for a cumulative generation rate of 20 Gb/s. The system relies principally on chip-based optoelectronic components that could be integrated in a compact, economical package.
AI snapshot
Written from the paper’s full text.
TL;DR
The paper demonstrates a scalable parallel physical random number generator that spectrally slices amplified spontaneous emission from a single superluminescent LED into two independent channels, each producing 10 Gb/s random bits for a cumulative 20 Gb/s. After delayed XOR processing, each channel and the interleaved 20 Gb/s sequence pass all NIST statistical tests for randomness.
Key findings
- Two nonoverlapping spectral filters (2.2 nm bandwidth, centered at 1540 and 1555 nm) produced two statistically independent 10 Gb/s random bit streams from one SLED, for a cumulative rate of 20 Gb/s.
- Cross-correlation and mutual information between two 10^9-bit sequences showed low bias and undetectable interchannel correlation.
- Each channel showed statistically detectable deviations from ideal behavior, attributed to nonideal detector characteristics; XORing each signal with a copy delayed by as little as 26 bits produced sequences that consistently passed all NIST tests.
- The XORed 10^9-bit record passed all 188 NIST statistical tests, and an interleaved 20 Gb/s sequence also passed all requisite tests. The authors project that additional filters could increase the number of parallel wavelength channels to at least 20, for a cumulative rate in excess of 200 Gb/s.
Methods
- Experimental optoelectronic setup: a fiber-coupled SLED (Exalos 1520) generated broadband optical noise, which was separated into two nonoverlapping spectral slices using wavelength-division multiplexing filters and detected by a pair of 11 GHz photoreceivers.
- Each received analog signal was threshold detected by a clocked comparator at a 10 GHz clock rate, with bias tees adjusting logic thresholds to balance ones and zeros; two bit error rate testers recorded simultaneous synchronous bit streams.
- Independence was quantified using cross-correlation and mutual information as functions of relative delay for two 10^9-bit sequences; randomness quality was tested using the NIST statistical test suite, including after delayed XOR processing.
Limitations
- The paper states that each channel exhibited statistically detectable deviations from ideal behavior, attributed to nonideal detector characteristics including the temporal response of the photoreceiver electronics, which can produce small correlations between successive bits.
- The delayed XOR process used to improve statistical properties was computed offline in this demonstration, though the authors state it could be implemented using real-time hardware.
- The scaling to at least 20 channels and >200 Gb/s is projected with additional filters, not demonstrated in the reported experiment.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
0 of 10 cited sources checked · 0 supported · 0 flagged
- Findingcitation§ Scalable parallel physical random number generator based on a superluminescent LED · p. 2
Delays as short as 26 bits are sufficient to produce sequences that consistently pass all of the NIST statistical tests for randomness.
“Delays as short as 26 bits were found to be sufficient to produce sequences that consistently pass all of the National Institute of Standards and Technology (NIST) statistical tests for randomness [17].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [17]
Not checked in the free run.
- Findingnone given§ Scalable parallel physical random number generator based on a superluminescent LED · p. 3
The system uses only commercially available optoelectronic components.
“The system uses only commercially available optoelectronic components that could be integrated at the chip- or board-level for high-performance random number generation in computing applications.”
Quote in the paperSubstantial claim - Findingown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1
Using a pair of nonoverlapping spectral filters and a fiber-coupled superluminescent LED, the authors produced two independent 10 Gb/s random bit streams, for a cumulative generation rate of 20 Gb/s.
“Using a pair of nonoverlapping spectral filters and a fiber-coupled superluminescent LED (SLED), we produced two independent 10 Gb = s random bit streams, for a cumulative generation rate of 20 Gb = s .”
- Also stated in: § Scalable parallel physical random number generator based on a superluminescent LED · p. 1 · boosted
“Specifically, we demonstrate simultaneous generation of two statistically independent 10 Gb = s random bit streams using spectrally distinct signals obtained from a single superluminescent LED (SLED).”
- Also stated in: § Scalable parallel physical random number generator based on a superluminescent LED · p. 3 · plain
“The system uses spectrally sliced light from a SLED to generate two parallel channels of random bits at 10 Gb = s each.”
Quote in the paperSubstantial claim - Also stated in: § Scalable parallel physical random number generator based on a superluminescent LED · p. 1 · boosted
- Findingown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 2
A 20 Gb/s sequence constructed by sequentially interleaving the two binary sequences (after XOR processing) passes all of the requisite statistical tests.
“We also constructed a 20 Gb = s sequence by sequentially interleaving the two binary sequences (after XOR processing). The interleaved sequence was also observed to pass all of the requisite statistical tests”
Quote in the paperSubstantial claim - Findingown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1
The spectrally sliced amplified spontaneous emission from the SLED produces a fast, fluctuating electrical signal that is much stronger than the background electrical noise.
“The spectrally sliced ASE produces a fast, fluctuating electrical signal that is much stronger than the background electrical noise.”
Quote in the paperSubstantial claim
- Findingown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 2
When each channel is subjected to more extensive statistical testing, statistically detectable deviations from ideal behavior are observed.
“Although the parallel binary sequences exhibit low bias and undetectable interchannel correlation, when each channel is subjected to more extensive statistical testing, we observed statistically detectable deviations from ideal behavior.”
Quote in the paperSubstantial claim - Interpretationown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1
This Letter describes the first demonstration that multiple independent streams can be simultaneously obtained from a single optoelectronic noise source without the need for external optical amplification or gain.
“While spontaneous emission has been recently exploited for random number generation [9], this Letter describes the first demonstration, to our knowledge, that multiple independent streams can be simultaneously obtained from a single optoelectronic noise source without the need for external optical amplification or gain.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [9] Williams, 2010, Opt. Express source
Not checked in the free run.
- Interpretationown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 3
The reported system is a scalable high-speed physical parallel random number generator based on amplified spontaneous emission from a single optoelectronic light source.
“In summary, we report a scalable high-speed physical parallel random number generator based on ASE from a single optoelectronic light source.”
Quote in the paperSubstantial claim - Interpretationown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 2
The interleaved sequence passing all requisite statistical tests further confirms the independence of the two sequences.
“which further confirms the independence of the sequences [18].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [18]
Not checked in the free run.
- Premisecitation§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1
With the exception of spatially multiplexed systems based on two-dimensional imaging, high-speed optical random number generators produce serial binary sequences that cannot be easily scaled to higher rates.
“With the exception of spatially multiplexed systems based on two-dimensional imaging [14,15], high-speed optical random number generators produce serial binary sequences that cannot be easily scaled to higher rates.”
Quote in the paperSubstantial claim0 of 2 sources checked - Interpretationown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 2
The observed deviations from ideal behavior are attributable to nonideal detector characteristics, including the temporal response of the photoreceiver electronics, which can produce small correlations between successive bits in each channel.
“We attribute these deviations to nonideal detector characteristics, including the temporal response of the photoreceiver electronics, which can produce small correlations between successive bits in each channel.”
Quote in the paperSubstantial claim - Premisecitation§ Scalable parallel physical random number generator based on a superluminescent LED · p. 2
Constructing the XOR between each signal and a time-delayed copy of itself significantly improves the statistical properties of each channel.
“As we previously demonstrated [9], the statistical properties of each channel can be significantly improved by constructing the XOR between each signal and a time-delayed copy of itself.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [9] Williams, 2010, Opt. Express source
Not checked in the free run.
- Interpretationnone given§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1
The system relies principally on chip-based optoelectronic components that could be integrated in a compact, economical package.
“The system relies principally on chip-based optoelectronic components that could be integrated in a compact, economical package.”
- Also stated in: § Scalable parallel physical random number generator based on a superluminescent LED · p. 3 · hedged
“The system uses only commercially available optoelectronic components that could be integrated at the chip- or board-level for high-performance random number generation in computing applications.”
Quote in the paperNot a substantial claim - Also stated in: § Scalable parallel physical random number generator based on a superluminescent LED · p. 3 · hedged
- May combine two claims. 3
- The quote only partly states this claim. 1
- Classified as probably not a substantial claim. 5
May combine two claims. The quote only partly states this claim.
Findingown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 2The XORed data set x[n] ⊕ x[n-26] passes all of the NIST statistical tests, and similar results are obtained for the second parallel channel.
“The XORed data set passes all of the NIST statistical tests, and we obtain similar results for the second parallel channel.”
Quote in the paperSubstantial claimMay combine two claims.
Findingnone given§ Scalable parallel physical random number generator based on a superluminescent LED · p. 2The parallel binary sequences exhibit low bias and undetectable interchannel correlation.
“Although the parallel binary sequences exhibit low bias and undetectable interchannel correlation, when each channel is subjected to more extensive statistical testing, we observed statistically detectable deviations from ideal behavior.”
Quote in the paperSubstantial claimClassified as probably not a substantial claim.
Premisecitation§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1Parallel random bit generation methods could dramatically improve the generation rate and scalability by producing many random bits simultaneously.
“Parallel random bit generation methods could dramatically improve the generation rate and scalability by producing many random bits simultaneously.”
Quote in the paperSubstantial claim0 of 4 sources checked- cites [10] Mascagni, 2000, ACM Trans. Math. Softw. source
Not checked in the free run.
- cites [11] Wolfram, 1986, Adv. Appl. Math. source
Not checked in the free run.
- cites [12] Hortensius, 1989, IEEE Trans. Comput. source
Not checked in the free run.
- cites [13] Ackermann, 2001, Comput. Phys. Commun. source
Not checked in the free run.
Probably not a substantial claim: may not pass S4 (based). May combine two claims.
Interpretationreasoning§ Scalable parallel physical random number generator based on a superluminescent LED · p. 3With additional filters, the number of parallel wavelength channels could potentially be increased to at least 20, for a cumulative generation rate in excess of 200 Gb/s using a single SLED source.
“With additional filters, the number of parallel wavelength channels could potentially be increased to at least 20, for a cumulative generation rate in excess of 200 Gb = s using a single SLED source.”
Quote in the paperNot a substantial claim- Rests on an assumption: The SLED's optical spectrum is broad enough to support at least 20 nonoverlapping wavelength channels.
Probably not a substantial claim: may not pass S3 (load-bearing).
Study design and conductreasoning§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1Both bit error rate testers acquire simultaneous and synchronous records.
“An external 10 GHz clock signal determines the bit generation rate, and a second external trigger signal is used to initiate the data acquisition process, ensuring that both BERTs acquire simultaneous and synchronous records.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
Study design and conductown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1The quality of each stream was tested using statistical tests developed for cryptographic security.
“We quantify the independence of the two streams by statistically computing the cross correlation and mutual information between them, and we test the quality of each stream using statistical tests developed for cryptographic security.”
Quote in the paperSubstantial claimClassified as probably not a substantial claim.
Study design and conductown data§ Scalable parallel physical random number generator based on a superluminescent LED · p. 1The independence of the two streams was quantified by statistically computing the cross correlation and mutual information between them.
“We quantify the independence of the two streams by statistically computing the cross correlation and mutual information between them, and we test the quality of each stream using statistical tests developed for cryptographic security.”
- Also stated in: § Scalable parallel physical random number generator based on a superluminescent LED · p. 2 · plain
“To better quantify the independence of the two sequences, we statistically computed the correlation ρ between the two binary sequences x ½ n Š and y ½ n þ k Š as a function of their relative delay k :”
Quote in the paperSubstantial claim- Also stated in: § Scalable parallel physical random number generator based on a superluminescent LED · p. 2 · plain
The paper relies on this without support.
The SLED's optical spectrum is broad enough to support at least 20 nonoverlapping wavelength channels.
“With additional filters, the number of parallel wavelength channels could potentially be increased to at least 20, for a cumulative generation rate in excess of 200 Gb = s using a single SLED source.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
To overcome the bandwidth limitation of shot noise-limited homodyne detection, researchers have developed QRNGs based on measuring phase or intensity noise of amplified spontaneous emission (ASE), which is quantum mechanical by nature.
“To overcome the bandwidth limitation of shot noise-limited homodyne detection, researchers have developed QRNGs based on measuring phase 40 -45 or intensity noise 46,47 of ampli fi ed spontaneous emission(ASE), which is quantum mechanical by nature. 15,48,49”
From Quantum random number generation (2016) · cites this paper as [47]
The citation lists haven’t been fetched yet
References 18 · Cited by 146
- Authors
- Li, X., Cohen, A. B., Murphy, T. E., & Roy, R.
- Affiliations
- University of Maryland, College Park (US)
- Venue
- Optics Letters
- Publisher
- Optica Publishing Group
- Published
- 2011-03-15
- Volume
- 36
- Issue
- 6
- Pages
- 1020
- ISSN
- 0146-9592 (ISSN-L)
- 1539-4794
- OpenAlex
- W2065883977
- Semantic Scholar
- 94498ab2003bc61f068e4b866b728da250316e0e
- Cited by
- 146
- References
- 18
- Citations by year
- 2026: 2 · 2025: 6 · 2024: 12 · 2023: 7 · 2022: 2 · 2021: 11 · 2020: 14 · 2019: 12 · 2018: 5 · 2017: 14 · 2016: 8 · 2015: 11 · 2014: 17 · 2013: 10 · 2012: 11
- FWCI
- 12.08
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Optics Letters (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- China Scholarship Council
- U.S. Department of Defense Multidisciplinary University Research Initiative (ONR N000140710734)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Quantum Information and Cryptography
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Photonic Communication Systems
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Semiconductor Lasers and Optical Devices
- Keywords
- physical random number generation, superluminescent diode, spectral filtering
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-09.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
