CoChat
Sign Up
Cited by 459Advanced Memory and Neural ComputingPhysical Unclonable Functions (PUFs) and Hardware SecurityChaos-based Image/Signal Encryption

A novel true random number generator based on a stochastic diffusive memristor

Open access
Jiang, Hao, Belkin, Daniel, Savel’ev, Sergey E., Lin, Siyan, Wang, Zhongrui, Li, Yunning, Joshi, Saumil, Midya, Rivu +6 more·2017·Nature Communications
DOI 10.1038/s41467-017-00869-xPMC PMC5638922
Open this paper in CoChat — free accountOpen PDF

Abstract

The intrinsic variability of switching behavior in memristors has been a major obstacle to their adoption as the next generation of universal memory. On the other hand, this natural stochasticity can be valuable for hardware security applications. Here we propose and demonstrate a novel true random number generator utilizing the stochastic delay time of threshold switching in a Ag:SiO 2 diffusive memristor, which exhibits evident advantages in scalability, circuit complexity, and power consumption. The random bits generated by the diffusive memristor true random number generator pass all 15 NIST randomness tests without any post-processing, a first for memristive-switching true random number generators. Based on nanoparticle dynamic simulation and analytical estimates, we attribute the stochasticity in delay time to the probabilistic process by which Ag particles detach from a Ag reservoir. This work paves the way for memristors in hardware security applications for the era of the Internet of Things.

AI snapshot

Written from the paper’s full text.

TL;DR

The paper demonstrates a true random number generator (TRNG) that uses the stochastic delay time of threshold switching in a Ag:SiO₂ diffusive memristor. Its generated bits pass all 15 NIST 800-22 randomness tests without post-processing, and the stochasticity is attributed to Ag particles detaching from an Ag reservoir.

Key findings

  • A Ag:SiO₂ diffusive memristor TRNG passed all 15 NIST Special Publication 800-22 randomness tests without any post-processing, described as a first for memristive-switching TRNGs; testing used 76 million binary bits divided into 76 sequences of 1 M bits each.
  • The 5 µm × 5 µm Pt/Ag/Ag:SiO₂/Pt device showed threshold switching at ~0.5 V, a >10⁵ ON/OFF window, sub-100 nA operation current, <pA OFF-state leakage, and self-OFF relaxation within 100 µs; at 0.5 V and 300 µs pulse width, a typical delay time was ~130 µs.
  • The proof-of-concept TRNG generated 6 kb s⁻¹ using the 6 lowest-order bits from a 16-bit counter clocked at 11.0592 MHz under 1 kHz input pulses; at 85 °C it still passed NIST tests but only 3 lowest-order bits could be collected, reducing bitrate from 6 to 3 kb s⁻¹.
  • Nanoparticle dynamic simulations and analytical estimates attributed the stochastic delay to the probabilistic process by which Ag particles detach from a Ag reservoir before forming conduction channel(s); simulated and measured delay-time distributions were similar, with higher voltage giving shorter and narrower delay-time distributions.

Methods

  • Device fabrication used Si wafers with 100 nm thermal SiO₂, photolithography-patterned Pt bottom electrodes, RF co-sputtered Ag:SiO₂ (~17% Ag atomic ratio by XPS), and a 5 nm Ag reservoir plus Pt/Au top electrode; electrical characterization used a Keysight B1500, waveform generator, and oscilloscope.
  • TRNG operation used a diffusive memristor, comparator, AND gate, and counter; a microcontroller (IAP15F2K61S2) with built-in 16-bit counter at 11.0592 MHz collected the 6 lower-order bits under 0.5 V, 300 µs pulses at 1 kHz, yielding 76 M bits for NIST 800-22 testing.
  • Simulations generalized a nanoparticle dynamics model with Langevin equations for Ag nanoparticles, heat dynamics, and memristor capacitance; 48 volatile switching cycles were simulated and compared with experimental delay-time distributions.

Limitations

  • Stated: pulse parameters must be optimized to avoid failure modes such as no switching to the low-resistance state or insufficient relaxation between pulses; higher-order counter bits require a higher clock frequency, increasing power consumption.
  • Stated: at 85 °C only the 3 lowest-order bits could be collected, reducing bitrate from 6 to 3 kb s⁻¹; a single device failed and stuck at ON after 54 M bits, corresponding to endurance of ~10⁷ cycles.
  • Evident from the text: the demonstration is a proof-of-concept breadboard circuit using a single device, with a current bitrate of 6 kb s⁻¹ that the authors describe as sufficient for some encryption applications but not all; long-term reliability is bounded by the observed device failure after ~10⁷ cycles.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41467-017-00869-x

Cites

40 of 40 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 42 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    When the applied voltage was removed, the device relaxed to the OFF state within 100 μs, as read by a subsequent 50 mV pulse.

    “When the applied voltage was removed, the device relaxed to the OFF state within 100 μs, as read by a subsequent 50 mV pulse.”
    Quote in the paperNot a substantial claim
  • Findingown data

    The Ag:SiO2 device had an extremely low (<pA) leakage current at OFF state.

    “The Ag:SiO2 device did not require electroforming and exhibited reliable threshold under quasi-DC sweeps with a > 105 ON/OFF window, a sub-100 nA operation current and an extremely low (< pA) leakage current at OFF state (Fig. 1b).”
    Quote in the paperNot a substantial claim
  • Findingreasoning

    Due to the stochastic nature of the delay time for each cycle, the counter output after each pulse was totally unpredictable and flipped randomly between '0' and '1'.

    “Due to the stochastic nature of the delay time for each cycle, the counter output after each pulse was totally unpredictable and flipped randomly between “0” and “1”.”
    Quote in the paperSubstantial claim
  • Findingown data

    The Ag:SiO2 device exhibited reliable threshold switching under quasi-DC sweeps with a >10^5 ON/OFF window.

    “The Ag:SiO2 device did not require electroforming and exhibited reliable threshold under quasi-DC sweeps with a > 105 ON/OFF window, a sub-100 nA operation current and an extremely low (< pA) leakage current at OFF state (Fig. 1b).”
    Quote in the paperSubstantial claim
  • Findingown data

    Under a 300 μs pulse of 0.5 V, a finite delay time (incubation period, ~130 μs) was required before Vout abruptly increased, indicating ON-switching.

    “Under this specific applied V in, a finite delay time (incubation period, ~ 130 µs) was required before V out abruptly increased, indicating ON-switching of the device.”
    Quote in the paperSubstantial claim
  • May combine two claims. 24
  • Classified as probably not a substantial claim. 17
  • The quote only partly states this claim. 1

  • May combine two claims.

    Findingown data

    A series resistor was effective in limiting the ON-state current and tuning the ON/OFF window.

    “A series resistor was found effective in limiting the ON-state current and tuning the ON/OFF window (Supplementary Fig. 2).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The device abruptly reached a low resistance state at a threshold voltage of around 0.5 V (ON-switching), followed by spontaneous relaxation back to the high-resistance state when the voltage swept back to below 0.3 V (self-OFF-switching).

    “The device abruptly reached a low resistance state at a threshold voltage of around 0.5 V (ON-switching), followed by a spontaneous relaxation back to the high-resistance state when the voltage swept back to below 0.3 V (self-OFF-switching), confirming the volatility of the device.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The fitted probability distributions appear to be consistent with experimental results, confirming the feasibility of the proposed mechanism.

    “The fitted probability distributions appear to be consistent with experimental results, confirming the feasibility of our proposed mechanism”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The diffusive memristor switches to a low-resistance state under a voltage pulse after a random delay time and relaxes back to the high-resistance state spontaneously upon removal of the applied electrical bias.

    “The device switches to a low-resistance state under a voltage pulse after a random delay time, and relaxes back to the high-resistance state spontaneously upon removal of the applied electrical bias.”
    • Also stated in: § Stochastic volatile switching behavior of a Ag:SiO2 diffusive memristor · plain
      “The device abruptly reached a low resistance state at a threshold voltage of around 0.5 V (ON-switching), followed by a spontaneous relaxation back to the high-resistance state when the voltage swept back to below 0.3 V (self-OFF-switching), confirming the volatility of the device.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The delay time statistics from simulations confirm a shorter delay time and a narrower distribution with increasing applied voltage amplitudes, similar to experimental results.

    “The delay time statistics from simulations also confirm a shorter delay time and a narrower distribution with increasing applied voltage amplitudes (Supplementary Fig. 5a), similar to experimental results shown in Fig. 1d.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    Each input pulse gives 6 random binary bits, for a total bit generation rate of 6 kbs−1.

    “Each input pulse will give us 6 random binary bits, for a total bit generation rate of 6 kbs−1.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    There is very good agreement for the simulated distribution of delay time and RC deterministic time delays t0, justifying good agreement between experimental and simulated data.

    “Similarly, we obtain a very good agreement for the simulated distribution of delay time, and RC deterministic time delays t 0, thus, justifying a good agreement between experimental and simulated data (Supplementary Fig. 5b).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Supplementary Fig. 5b

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Figure 4i shows the relationship between t0 and applied voltages with a curve fit based on equation (3).

    “Figure 4g, h show the fitting results for the distribution of delay time under 0.4 and 0.7 V from experiments based on eq. (2) while Fig. 4i shows the relationship between t 0 and applied voltages with a curve fit based on eq. (3).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Dense Ag nanoclusters (mostly 2 to 5 nm in diameter with a few outliers of 10 nm) are uniformly dispersed in the SiO2 matrix.

    “dense Ag nanoclusters (mostly 2 to 5 nm in diameter with a few outliers of 10 nm) were uniformly dispersed in the SiO2 matrix (Supplementary Fig. 1b).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    A higher voltage leads to a shorter average delay time with a narrower distribution.

    “A higher voltage leads to a shorter average delay time with a narrower distribution.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingown data

    Figure 4g,h show fitting results for the distribution of delay time under 0.4 and 0.7 V from experiments based on equation (2).

    “Figure 4g, h show the fitting results for the distribution of delay time under 0.4 and 0.7 V from experiments based on eq. (2) while Fig. 4i shows the relationship between t 0 and applied voltages with a curve fit based on eq. (3).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Diffusion in higher dimensions (3D) does not qualitatively change the described simple physical picture; the mechanism described is valid until the electric field is not applied perpendicular to the conducting paths.

    “We have also checked (Supplementary Note 3 and Supplementary Fig. 8) if diffusion in higher dimensions (3D) can qualitatively change the described above simple physical picture and concluded that the mechanism described is valid until the electric field is not applied perpendicular the conducting paths.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Supplementary Note 3

      Not checked in the free run.

    • cites Supplementary Fig. 8

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    We proposed and experimentally demonstrated a novel diffusive memristor true random number generator (diffusive memristor TRNG) utilizing the stochastic delay time as the source of randomness.

    “In conclusion, we proposed and experimentally demonstrated a novel diffusive memristor true random number generator (diffusive memristor TRNG) utilizing the stochastic delay time as the source of randomness.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingown data

    Supplementary Fig. 10 shows the comparison between frequency counts of 8-bit block values if the 8-lower-order bits each cycle are collected and those of 6-bit block values from only 6-lowest-order bits.

    “Supplementary Fig. 10 shows the comparison between frequency counts of 8-bit block values if the 8-lower-order bits each cycle are collected and those of 6-bit block values from only 6-lowest-order bits.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Supplementary Fig. 10

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    At 85 °C, we can only collect the 3 lowest-order bits, which means the bitrate decreases from 6 to 3 kb s−1.

    “Unlike in operation at room temperature, we can only collect the 3 lowest-order bits, which means the bitrate decreases from 6 to 3 kb s−1. at 85 °C.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Binary bit sequences generated by our diffusive memristor TRNG passed all the 15 NIST Special Publication 800-22 randomness tests without any post-processing, a first for hardware utilizing memristive switching.

    “Binary bit sequences generated by our diffusive memristor TRNG passed all the 15 NIST Special Publication 800-22 randomness tests without any post-processing, a first for hardware utilizing memristive switching.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Our diffusive memristor TRNG still functions satisfactorily and passes the NIST tests even at 85 °C.

    “Our diffusive memristor TRNG still functions satisfactorily and passes the NIST tests even at 85 °C, as shown in the Supplementary Table 2.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Supplementary Table 2

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    We collected 11 M binary bits under 1 kHz pulses (voltage amplitude: 0.5 V and pulse width: 300 μs).

    “We collected 11 M binary bits under 1 kHz pulses (voltage amplitude: 0.5 V and pulse width: 300 µs).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    We collected 54 M binary bits from a single diffusive memristor before the device failed and got stuck at ON state (each cycle produced 6 random bits with endurance of ~10^7 cycles).

    “We collected 54 M binary bits from a single diffusive memristor before the device failed and got stuck at ON state (each cycle produced 6 random bits with endurance of ~ 107 cycles).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    We performed simulation with MATLAB and demonstrated the bitrate can be readily increased by 50 times (to 300 kb s−1) with this method.

    “We performed simulation with MATLAB and demonstrated the bitrate can be readily increased by 50 times (to 300 kb s−1) with this method (Supplementary Note 4, Supplementary Fig. 11 and Supplementary Table 1).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Supplementary Note 4

      Not checked in the free run.

    • cites Supplementary Fig. 11

      Not checked in the free run.

    • cites Supplementary Table 1

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    We simulated three possible cases during continuous pulse switching and identified two ways in which random number generation could fail: (i) no switching to the low resistive state during pulse, or (ii) insufficient time to relax to the high resistive state during the inter-pulse interval.

    “We simulated three possible cases during continuous pulse switching and identified two ways in which random number generation could fail: (i) no switching to the low resistive state during pulse, or (ii) insufficient time to relax to the high resistive state during the inter-pulse interval (Supplementary Fig. 9).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Supplementary Fig. 9

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Frequency counts of 6-bit block values are uniform but those for the 8-bit block values are not.

    “Clearly, frequency counts of 6-bit block values are uniform but those for the 8-bit block values are not”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    The time needed to charge the capacitor is deterministic and delays setting voltage across the device, while the time for Ag transportation is very short and in practice can be neglected.

    “The time needed to charge the capacitor (i) is deterministic in nature and delays setting voltage across the device, while the time for Ag transportation (iii) is very short (from spot 2 to spot 4, due to a fast increase in device temperature and, thus, very fast diffusion) and in practice can be neglected.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationcitation

    The diffusive memristor TRNG can easily be incorporated into memory subsystems, greatly increasing security and area efficiency.

    “The diffusive memristor TRNG can easily be incorporated into memory subsystems, greatly increasing the security, and the area efficiency33.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [33] Wei, Z. et al. True random number generator using current difference based on a fractional stochastic model in 40-nm embedded ReRAM. IEEE Electron. Dev. Meet. 4.8.1–4.8.4 (San Franciso, CA, USA, 2016). source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The bit on which the counter stops is random because the stochastic delay time leads to random pulse width from the comparator and thus a random number of clock pulses sent to the counter.

    “The bit on which the counter stops is random, because the stochastic delay time of the diffusive memristor leads to random pulse width from the comparator (V 3) and thus a random number of clock pulses that are sent to the counter.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Possible degradation due to long pulse cycling is the other concern for memristive switching based TRNG.

    “Possible degradation due to long pulse cycling is the other concern for memristive switching based TRNG39.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [39] Gallo, M. L., Tuma, T., Zipoli, F., Sebastian, A., Eleftheriou, E. Inherent stochasticity in phase-change memory devices. in 46th European Solid State Device Research Conference (ESSDERC), 373–376 (Lausanne, Switzerland, 2016). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    A higher frequency clock signal comes at the cost of increased power consumption.

    “This comes at the cost of increased power consumption.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    As a result, simpler circuits are required.

    “As a result, simpler circuits are required.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    The decreased bitrate at high temperatures could be a result of the decreased s.d. of delay time at high temperatures.

    “and so the decreased bitrate could be a result of the decreased s.d. of delay time at high temperatures.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisenone given

    In previous threshold-value approaches, the circuit outputs 1 if the measured value exceeds the threshold, and 0 otherwise.

    “The circuit will output 1 if the measured value exceeds the threshold, and 0 otherwise.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning

    Our diffusive memristor TRNG can be built into memory subsystems, reducing chip area and increasing security.

    “Our diffusive memristor TRNG can be built into memory subsystems, reducing chip area and increasing security.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    Randomness from volatile switching with a high ON/OFF ratio is easier to exploit than small noises or current fluctuations.

    “Randomness from volatile switching with a high ON/OFF ratio is easier to exploit than small noises or current fluctuations.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). The quote only partly states this claim.

    Interpretationnone given

    The high frequency clock used to generate random numbers makes sure the randomness is not affected by temperature-induced shifts.

    “However, the high frequency clock used to generate random numbers makes sure the randomness is not affected.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    To further increase the bit rate for other applications, a counter with more bits can be used in the circuit.

    “To further increase the bit rate for other applications, a counter with more bits can be used in the circuit.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationnone given

    Parallel operation of several diffusive memristors could also lead to increased bitrates by simultaneously generating multiple random sequences.

    “Finally, parallel operation of several diffusive memristors could also lead to increased bitrates by simultaneously generating multiple random sequences.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The physical nature of stochastic delay time during ON-switching in these devices was explained by nanoparticle dynamics simulation and simple analytical estimates and attributed to the ionic/atomic motion process.

    “Finally, the physical nature of stochastic delay time during ON-switching in these devices was explained by nanoparticle dynamics simulation and simple analytical estimates and attributed to the ionic/atomic motion process”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    NbO2 has sub-nanosecond switching speed, albeit with relatively high operation current (~300 μA) and voltage (~1.5 V) even with small device size.

    “For instance, NbO2 could be a good candidate for the TRNG because of its sub-nanosecond switching speed, albeit with relatively high operation current ( ~ 300 µA) and voltage ( ~ 1.5 V) even with small device size20.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [20] Pickett, M. D. & Williams, R. S. Sub-100 fJ and sub-nanosecond thermally driven threshold switching in niobium oxide crosspoint nanodevices. Nanotechnology. 23, 215202 (2012). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationnone given

    Device engineering that leads to a higher switching speed of the diffusive memristors will also improve the bitrate without changing the TRNG circuit.

    “In addition to circuit solutions, device engineering that leads to a higher switching speed of the diffusive memristors will also improve the bitrate without changing the TRNG circuit.”
    Quote in the paperNot a substantial claim
38 of 38

  • The authors state this as an assumption.

    The particular shape of the potential profile used in simulations does not significantly affect the result.

    “The particular shape of potential (the potential profile used in simulations is shown in Fig. 4f) does not significantly affect the result.”

  • The authors state this as an assumption.

    The resistance of the memristor has a tunneling nature.

    “We assume the resistance has a tunneling nature and is described by the equation where x 0 and x N+1 are positions of the device terminals, R t resistance amplitude and λ is the tunneling length.”

  • The authors state this as an assumption.

    The memristor capacitance C_M is a constant and is not a function of Ag nanoparticle locations.

    “In general, this capacitance could be different in the two memristor states37, 38, but we, for simplicity, assume that C M is a constant and is not a function Ag nano-particle locations.”

  • The paper relies on this without support.

    The actual system temperature and the macroscopic-cluster temperature can be significantly different.

    “Note that the actual system temperature and the macroscopic-cluster temperature can be significantly different.”

  • The authors state this as an assumption.

    The only important property of the potential is the large minima associated with interfacial energy compared with both temperature and depths of multiwell pinning potentials; the repulsive potential barrier of the Pt-electrode should be strong enough to ensure relaxation to high resistive state on a reasonable time scale when no voltage is applied.

    “The only important property of the potential is the large minima associated with interfacial energy comparing with both temperature and depths of multiwell pinning potentials; the repulsive potential barrier of the Pt-electrode should be strong enough to ensure relaxation to high resistive state on a reasonable time scale when no voltage is applied.”
5 of 5

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?

  • Their lack of compatibility with CMOS fabrication standards makes it difficult to apply them directly to the current CMOS logic platform used for mass production.

    “Moreover, their lack of compatibility with CMOS fabrication standards makes it difficult to apply them directly to the current CMOS logic platform used for mass production.”

    From Cryptographic transistor for true random number generator with low power consumption (2024) · cites this paper as [13]

    The Introduction (page unavailable) lists CMOS compatibility among the attractive properties of memristors, contrary to the claim that they lack it. The paper does not say that incompatibility prevents their direct use in mass-produced CMOS logic. (LLM, conf=0.87)

    Checked by a reasoning model — may be wrong.

    Experimental — accuracy not yet measured

  • tRNGs such as RRAM or MRAM are widely studied given their inherent switching stochasticity as an entropy source, which relies on material-induced variations.

    “For example, tRNGs such as RRAM or MRAM, which are not Si-based complementary metal-oxide semiconductor (CMOS) devices, are widely studied given their inherent switching stochasticity as an entropy source, which relies on material-induced variations (11–18).”

    From Cryptographic transistor for true random number generator with low power consumption (2024) · cites this paper as [13]

    The Introduction (page unavailable) describes memristor-based TRNGs that use intrinsic switching variation, and the Discussion attributes this paper’s stochasticity to ionic motion. These passages support the RRAM-related part of the claim, but our copy does not address MRAM-based TRNGs or establish the claimed distinction from Si-based CMOS devices. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • These tRNGs require a few entropy extraction circuits, each of which consumes power at the milliwatt level.

    “However, these tRNGs require a few entropy extraction circuits, each of which consumes power at the milliwatt level.”

    From Cryptographic transistor for true random number generator with low power consumption (2024) · cites this paper as [13]

    The Introduction and Discussion (pages unavailable) say that some earlier TRNG approaches needed sophisticated circuits or post-processing to improve randomness. They do not say that these TRNGs require a few entropy-extraction circuits, or that each circuit consumes power at the milliwatt level. (LLM, conf=0.78)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Random excursions lacks a P-value despite not being listed as an exception

    Minor

    The paper says P-values are omitted only for the non-overlapping-template and random excursions variant tests, but Table 1 also omits the P-value for random excursions.

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 1

  1. Liu, N., Pinckney, N., Hansen, S., Sylvester, D. & Blaauw, D. 2011 Symposium on VLSI Circuits 216–217 (Honolulu, HI, USA, 2011).

Cited by

Sorted by citations

Related

Authors
Jiang, H., Belkin, D., Savel’ev, S. E., Lin, S., Wang, Z., Li, Y., Joshi, S., Midya, R., Li, C., Rao, M., Barnell, M., Wu, Q., Yang, J. J., & Xia, Q.
Affiliations
  • University of Massachusetts Amherst (US)
  • Swarthmore College (US)
  • U.S. Air Force Research Laboratory Information Directorate (US)
Corresponding author
  • Yang, J. Joshua
  • Xia, Qiangfei
ORCID
Venue
Nature Communications
Publisher
Springer Science and Business Media LLC
Published
2017-10-12
Volume
8
Issue
1
Pages
882-882
ISSN
2041-1723 (ISSN-L)
DOI
10.1038/s41467-017-00869-x
PMC
PMC5638922
OpenAlex
W2762731122
Semantic Scholar
718b82a08c124bf96cbf9606752970a54dc5270b
Cited by
459
References
40
Citations by year
2026: 37 · 2025: 54 · 2024: 68 · 2023: 61 · 2022: 59 · 2021: 57 · 2020: 60 · 2019: 39 · 2018: 21 · 2017: 2 · 2012: 1
FWCI
13.85
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • Loughborough University Institutional Repository (Loughborough University) (submittedVersion, cc-by)
  • Loughborough University Research Repository (Loughborough University) (submittedVersion, cc-by)
Funders
  • National Science Foundation (1253073)
  • U.S. Air Force (FA9550, FA9550-12-1)
  • Division of Electrical, Communications and Cyber Systems
  • Air Force Office of Scientific Research (FA9550-, -15-2-, FA9550-12, FA9550‐12‐1‐0038, FA9550‐12‐1)
  • Air Force Research Laboratory (FA8750-15-2- 0044, FA9550)
Topics
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Memory and Neural Computing
  • Physical Sciences › Computer Science › Hardware and Architecture › Physical Unclonable Functions (PUFs) and Hardware Security
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Chaos-based Image/Signal Encryption
Keywords
true random number generator, memristors, diffusive memristors, threshold switching, silver nanoparticles, NIST randomness tests, hardware security, Internet of Things, memristive switching

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