Intrinsic Thermal Noise Research Articles

Thermal noise-driven resonant sensors

MEMS/NEMS resonant sensors hold promise for minute mass and force sensing. However, one major challenge is that conventional externally driven sensors inevitably encounter undesired intrinsic noise, which imposes a fundamental limitation upon their signal-to-noise ratio (SNR) and, consequently, the resolution. Particularly, this restriction becomes increasingly pronounced as sensors shrink to the nanoscale. In this work, we propose a counterintuitive paradigm shift that turns intrinsic thermal noise from an impediment to a constituent of the sensor by harvesting it as the driving force, obviating the need for external actuation and realizing ‘noise-driven’ sensors. Those sensors employ the dynamically amplified response to thermal noise at resonances for stimulus detection. We demonstrate that lightly damped and highly compliant nano-structures with high aspect ratios are promising candidates for this class of sensors. To overcome the phase incoherence of the drive force, three noise-enabled quantitative sensing mechanisms are developed. We validated our sensor paradigm by experimental demonstrating noise-driven pressure and temperature sensors. Noise-driven sensors offer a new opportunity for delivering practical NEMS sensors that can function at room temperature and under ambient pressure, and a development that suggests a path to cheaper, simpler, and low-power-consumption sensors.

Entropic stochastic resonance of finite-size particles in confined Brownian transport.

We demonstrate the existence of entropic stochastic resonance (ESR) of passive Brownian particles with finite size in a double- or triple-circular confined cavity, and compare the similarities and differences of ESR in the double-circular cavity and triple-circular cavity. When the diffusion of Brownian particles is constrained to the double- or triple-circular cavity, the presence of irregular boundaries leads to entropic barriers. The interplay between the entropic barriers, a periodic input signal, the gravity of particles, and intrinsic thermal noise may give rise to a peak in the spectral amplification factor and therefore to the appearance of the ESR phenomenon. It is shown that ESR can occur in both a double-circular cavity and a triple-circular cavity, and by adjusting some parameters of the system, the response of the system can be optimized. The differences are that the spectral amplification factor in a triple-circular cavity is significantly larger than that in a double-circular cavity, and compared with the ESR in a double-circular cavity, the ESR effect in a triple-circular cavity occurs within a wider range of external force parameters. In addition, the strength of ESR also depends on the particle radius, and smaller particles can induce more obvious ESR, indicating that the size effect cannot be safely neglected. The ESR phenomenon usually occurs in small-scale systems where confinement and noise play an important role. Therefore, the mechanism that is found could be used to manipulate and control nanodevices and biomolecules.

Microseismic Observation Enabled by High-Sensitivity Micromechanical Interferometers

High-sensitivity accelerometry has historically been employed to explore Earth sciences and advance resource-related industries, such as petroleum exploitation. The devices capable of such measurements, however, traditionally involve precise mechanical structures and consequent large dimensions. In this work, combined by micromechanical systems and grating-assisted interferometry, a closed-loop optical accelerometer with nano-g (ng) sensitivity is developed for microseismic observations. In order to lower the intrinsic thermal noise, a modified 8.84 Hz micromechanical spring-mass oscillator is designed and fabricated. Additionally, an active control system is equipped to stabilize the device temperature and thus enhance the low-frequency stability. Field tests demonstrate that the prototype has achieved a scale factor of 1195 V/g, along with a dynamic range of 1.12 mg and a bandwidth of about 7 Hz. Further, under the controlled 45 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\pm$</tex-math></inline-formula> 0.01 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^\circ$</tex-math></inline-formula> C temperature condition, complete terrestrial microsesimic peaks from 0.1 Hz to 3 Hz are tracked. After calibration by a reference seismometer, the device self-noise is determined as 2 ng/ <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sqrt{Hz}$</tex-math></inline-formula> from 0.2 Hz to 4 Hz with a bias stability of 7.9 ng. The transducer and control circuits are both packaged within a 13.6 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\times 13.6\times$</tex-math></inline-formula> 12.8 cm <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$^{3}$</tex-math></inline-formula> insulated enclosure, making it advantageous for outdoor observations.

Quantum Parallel Training of a Boltzmann Machine on an Adiabatic Quantum Computer

AbstractDespite the anticipated speed‐up of quantum computing, the achievement of a measurable advantage remains subject to ongoing debate. Adiabatic Quantum Computers (AQCs) are quantum devices designed to solve quadratic uncostrained binary optimization (QUBO) problems, but their intrinsic thermal noise can be leveraged to train computationally demanding machine learning algorithms such as the Boltzmann Machine (BM). Despite an asymptotic advantage is expected only for large networks, a limited quantum speed up can be already achieved on a small BM is shown, by exploiting parallel adiabatic computation. This approach exhibits a 8.6‐fold improvement in wall time on the Bars and Stripes dataset when compared to a parallelized classical Gibbs sampling method, which has never been outperformed before by quantum approaches.

Intrinsic Temporal Performance of the RF Receive Coil in Magnetic Resonance Imaging.

The functional magnetic resonance imaging (fMRI) at ultra-high field (UHF, [Formula: see text]) is a powerful temporal acquisition method which promises to capture neuronal activities at submillimeter scale. But high-spatial-resolution fMRI still remains difficult, as the nuisance temporal noise which also grows with the main magnetic field strength. For decades, mainstream solutions in reducing motion-induced temporal noise include motion-correction algorithms in image post-processing as well as MR acquisition schemes in RF pulse sequence designs, however hardware-related studies have been rarely reported over the RF receive coil. In this study, we have proposed the intrinsic temporal performance model, which is specifically used for measuring coil-related intrinsic temporal SNR (tSNR*), and the intrinsic sensitivity variability and thermal noise variability have been proposed as model parameters. The intrinsic temporal performance of single-channel loops and array coils were evaluated using numerical electromagnetic simulations, and phantom experiments were designed to investigate the intrinsic thermal noise variability. It was observed that the achievable intrinsic tSNR* can be greatly lowered by ~90% even with 2 mm translational motion in the normal direction, suggesting the effect of RF receive coils in producing temporal noise. The proposed model provides a new perspective in optimizing coil designs and array coil temporal combination methods, which may offer a feasible means in achieving submillimeter resolutions at UHF. Moreover, model parameters from the intrinsic temporal performance model can be directly calculated based on single MRI acquisition, offering a practical performance metric for manufactures and customers in quality and assurance checks of RF receive coil products.

An energy-dependent electro-thermal response model of CUORE cryogenic calorimeter

The Cryogenic Underground Observatory for Rare Events (CUORE) is the most sensitive experiment searching for neutrinoless double-beta decay (0νββ) in 130Te. CUORE uses a cryogenic array of 988 TeO2 calorimeters operated at ∼10 mK with a total mass of 741 kg. To further increase the sensitivity, the detector response must be well understood. Here, we present a non-linear thermal model for the CUORE experiment on a detector-by-detector basis. We have examined both equilibrium and dynamic electro-thermal models of detectors by numerically fitting non-linear differential equations to the detector data of a subset of CUORE channels which are well characterized and representative of all channels. We demonstrate that the hot-electron effect and electric-field dependence of resistance in NTD-Ge thermistors alone are inadequate to describe our detectors' energy-dependent pulse shapes. We introduce an empirical second-order correction factor in the exponential temperature dependence of the thermistor, which produces excellent agreement with energy-dependent pulse shape data up to 6 MeV. We also present a noise analysis using the fitted thermal parameters and show that the intrinsic thermal noise is negligible compared to the observed noise for our detectors.

Quantum limits of position-sensitive photodiodes.

The split photodiode and the lateral effect photodiode are two popular detectors for measuring beam displacement. For small displacements of a Gaussian beam, which is the case of interest here, they are often seen as equivalent and used interchangeably, giving a signal proportional to the displacement. We show theoretically and experimentally that in the limit of low technical noise, where the signal to noise ratio is dominated by the shot noise of the light, the lateral effect photodiode produces a better signal to noise ratio than the split photodiode, owing to its optimum spatial detector response. This quantum advantage can be practically exploited in spite of the intrinsic thermal noise of the lateral effect photodiode.

A Mach-Zehnder Fabry-Perot hybrid fiber-optic interferometer operating at the thermal noise limit

A new type of interferometric fiber sensor based on a Mach-Zehnder Fabry-Perot hybrid scheme has been experimentally demonstrated. The interferometer combines the benefits of both a double-path configuration and an optical resonator, leading to record-high strain and phase resolutions limited only by the intrinsic thermal noise in optical fibers across a broad frequency range. Using only off-the-shelf components, the sensor is able to achieve noise-limited strain resolutions of 40 fvarepsilon /sqrt{(}Hz) at 10 Hz and 1 fvarepsilon /sqrt{(}Hz) at 100 kHz. With a proper scale-up, atto-strain resolutions are believed to be within reach in the ultrasonic frequency range with such interferometers.

Voltage-Controlled Domain Wall Motion-Based Neuron and Stochastic Magnetic Tunnel Junction Synapse for Neuromorphic Computing Applications

This work discusses the proposal of a spintronic neuromorphic system with spin orbit torque-driven domain wall motion (DWM)-based neurons and synapses. We propose a voltage-controlled magnetic anisotropy DWM-based magnetic tunnel junction (MTJ) neuron. We investigate how the electric field at the gate (pinning site), generated by the voltage signals from pre-neurons, modulates the DWM, which reflects in the nonlinear switching behavior of neuron magnetization. For the implementation of synaptic weights, we propose a 3-terminal MTJ with stochastic DWM in the free layer. We incorporate intrinsic pinning effects by creating triangular notches on the sides of the free layer. The pinning of the domain wall and intrinsic thermal noise of the device lead to the stochastic behavior of DWM. The control of this stochasticity by the spin orbit torque is shown to realize the potentiation and depression of the synaptic weight. The micromagnetics and spin transport studies in synapses and neurons are carried out by developing a coupled micromagnetic non-equilibrium Green’s function ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MuMag-NEGF</i> ) model. The minimization of the writing current pulsewidth by leveraging the thermal noise and demagnetization energy is also presented. Finally, we discuss the implementation of digit recognition by the proposed system using a spike time-dependent algorithm.

Bias-Dependent Intrinsic RF Thermal Noise Modeling and Characterization of Single-Layer Graphene FETs

In this article, the bias dependence of intrinsic channel thermal noise of single-layer (SL) graphene field-effect transistors (GFETs) is thoroughly investigated by experimental observations and compact modeling. The findings indicate an increase of the specific noise as drain current increases, whereas a saturation trend is observed at very high carrier density regime. Besides, short-channel effects, such as velocity saturation (VS) also result in an increment of noise at higher electric fields. The main goal of this work is to propose a physics-based compact model that accounts for and accurately predicts the above experimental observations in short-channel GFETs. In contrast to long-channel MOSFET-based models adopted previously to describe thermal noise in graphene devices without considering the degenerate nature of graphene, in this work, a model for short-channel GFETs embracing the 2-D material’s underlying physics and including a bias dependence is presented. The implemented model is validated with deembedded high-frequency data from two short-channel devices at quasi-static (QS) region of operation. The model precisely describes the experimental data for a wide range of low-to-high drain current values without the need of any fitting parameter. Moreover, the consideration of the degenerate nature of graphene reveals a significant decrease of noise in comparison with the nondegenerate case and the model accurately captures this behavior. This work can also be of utmost significance from the circuit designers’ aspect since noise excess factor, a very important figure of merit for RF circuits implementation, is defined and characterized for the first time in graphene transistors.

Low frequency 1/f noise in deep submicrometer-sized magnetic tunnel junctions

Magnetic tunnel junctions (MTJs) have been identified as promising candidates for the development of high-performance, ultra-low field magnetometers due to their high sensitivity, low cost, low power consumption, and small size. However, 1/f noise is often quite large at low frequencies and inevitably becomes one of the most difficult issues in developing a magnetic field sensor with ultrahigh sensitivity. Low-frequency 1/f noise can have both electric and magnetic origins, and it is a result of complex non-linear interactions between many degrees of freedom inside a sensor. Therefore, a reduction of the 1/f noise can be expected for the magnetic sensor with very small dimensions. Here, the dependence of the 1/f noise on voltage and strong hard-axis bias field in deep submicrometer-sized MgO-based MTJs is investigated with various junction sizes. The noise spectra were measured by using a home-built low-frequency noise measurement setup with maximum frequency up to 30 kHz. We find that the noise spectral power density is 1/f-like at low frequencies. The experimental results suggest a relative reduction of 1/f noise with respect to the intrinsic thermal noise in small-sized MTJs. The results may open a new approach for reducing the 1/f noise level in MTJ nanosensors.

Harnessing Dispersion in Soliton Microcombs to Mitigate Thermal Noise.

We explore intrinsic thermal noise in soliton microcombs, revealing thermodynamic correlations induced by nonlinearity and group-velocity dispersion. A suitable dispersion design gives rise to control over thermal-noise transduction from the environment to a soliton microcomb. We present simulations with the Lugiato-Lefever equation (LLE), including temperature as a stochastic variable. By systematically tuning the dispersion, we suppress repetition-rate frequency fluctuations by up to 50 decibels for different LLE soliton solutions. In an experiment, we observe a measurement-system-limited 15-decibel reduction in the repetition-rate phase noise for various settings of the pump-laser frequency, and our measurements agree with a thermal-noise model. Finally, we compare two octave-spanning soliton microcombs with similar optical spectra and offset frequencies, but with designed differences in dispersion. Remarkably, their thermal-noise-limited carrier-envelope-offset frequency linewidths are 1MHz and 100Hz, which demonstrates an unprecedented potential to mitigate thermal noise. Our results guide future soliton-microcomb design for low-noise applications, and, more generally, they illuminate emergent properties of nonlinear, multimode optical systems subject to intrinsic fluctuations.

Fundamental Thermal Limits on Data Retention in Low-Voltage CMOS Latches and SRAM

Ultra-low-power systems with substantial computing capacity require latches and SRAMs to operate at extremely low supply voltages. However, with aggressive technology scaling, reliability becomes a major challenge due to unavoidable process variations and the presence of multiple noise sources, including intrinsic thermal noise. This paper provides a quantitative measure of reliability by calculating the probability distribution function (PDF) of errors induced by thermal noise in latches and SRAMs operating in subthreshold conditions. Implemented in a novel simulation tool for thermal-noise analysis of CMOS circuits (STTACC), our algorithm uses a stochastic differential equation circuit model that preserves the proper Poissonian statistics for thermal-noise-driven current fluctuations in MOSFETs. Our probabilistic error model can handle error rate analysis for arrays of latches or full SRAMs on time scales from seconds to years without excessive computational overhead. We demonstrate that the time-to-error (TTE) statistics of subthreshold SRAMs obey log-normal distributions that depend on parameters such as node and device capacitance, device threshold variations and operating conditions of supply voltage and temperature. This makes it possible to quantitatively evaluate the asymptotic behavior of extremely rare error events that are inaccessible to standard SPICE-based simulations.

Experimental study on 1/f intrinsic thermal noise in optical fibers

Intrinsic thermal noise in optical fibers is an ultimate factor limiting the performances of fiber-based sensors and measurement systems. Therefore, it is important to have a thorough understanding of this kind of noise. However, the mechanism of the intrinsic thermal noise which has a 1/&lt;i&gt;f&lt;/i&gt; spectral density remains unclear so far. There are two theoretical models to explain the mechanism of this kind of noise: thermoconductive noise model and thermomechanical noise model. The thermoconductive noise model states that the noise is caused by diffusion of local entropy fluctuations associated with random spontaneous emission events, while the thermomechanical noise model says that the noise is caused by spontaneous fluctuations of fiber length induced by mechanical dissipation. Which theoretical model is correct is still an open question. In this paper, we experimentally investigate the intrinsic thermal noise in optical fibers by using a balanced fiber Michelson interferometer with heterodyne detection technique. When a fiber-stabilized laser with ultralow frequency noise is used as a laser source and other noise sources are carefully controlled, the 1/&lt;i&gt;f&lt;/i&gt; spectral intrinsic thermal noise can be observed down to infrasonic frequency. According to these measurements, in order to verify which theoretical model is the mechanism of generating the intrinsic thermal noise with 1/&lt;i&gt;f&lt;/i&gt; spectral density, we study the relationship between the level of the intrinsic thermal noise and the concentration of the dopant in fibers and the applied tension of fibers. We observe that the level of the 1/&lt;i&gt;f&lt;/i&gt; spectral intrinsic thermal noise is independent of the concentration of the dopant in fibers. This means that the thermoconductive noise model is not suitable to this case. We also observe that the level of the 1/&lt;i&gt;f&lt;/i&gt; spectral intrinsic thermal noise can be reduced by increasing the tension exerted on the optical fibers. Because the mechanical loss of a fiber can be lower than the loss of the material which the fiber is made of when the fiber is subjected to a certain tension, this observation proves the fact that the 1/&lt;i&gt;f&lt;/i&gt; spectral intrinsic thermal noise in optical fibers originates from the mechanical dissipation process inside optical fibers. This is consistent with the predictions of the thermomechanical noise model. Finally, the inconsistency between the experimental data and the theoretical results for thermomechanical noise is discussed.

Intrinsic mirror noise in Fabry\u2013Perot based polarimeters: the case for the measurement of vacuum magnetic birefringence

Although experimental efforts have been active for about 30 years, a direct laboratory observation of vacuum magnetic birefringence, due to vacuum fluctuations, still needs confirmation: the predicted birefringence of vacuum is Delta n = 4.0times 10^{-24} @ 1 T. Key ingredients of a polarimeter for detecting such a small birefringence are a long optical path within the magnetic field and a time dependent effect. To lengthen the optical path a Fabry–Perot is generally used with a finesse ranging from {{mathscr {F}}} approx 10^4 to {{mathscr {F}}} approx 7times 10^5. Interestingly, there is a difficulty in reaching the predicted shot noise limit of such polarimeters. We have measured the ellipticity and rotation noises along with Cotton-Mouton and Faraday effects as a function of the finesse of the cavity of the PVLAS polarimeter. The observations are consistent with the idea that the cavity mirrors generate a birefringence-dominated noise whose ellipticity is amplified by the cavity itself. The optical path difference sensitivity at 10;hbox {Hz} is S_{Delta {{mathscr {D}}}}=6times 10^{-19};hbox {m}/sqrt{mathrm{Hz}}, a value which we believe is consistent with an intrinsic thermal noise in the mirror coatings. Our findings prove that the continuous efforts to increase the finesse of the cavity to improve the sensitivity has reached a limit.

Achieving Ultra-Conformability With Polyimide-Based ECoG Arrays.

Micro-electrode arrays for electrocorticography (ECoG) represent the best compromise between invasiveness and signal quality, as they are surface devices that still allow high sensitivity recordings. In this work, an assessment of different technical aspects determining the ultimate performance of ultra-conformable polyimide-based μECoG arrays is conducted via a finite element model, impedance spectroscopy measurements and recordings of sensorimotor evoked potentials (SEPs) in rats. The finite element model proves that conformability of thin-film arrays can be achieved with polyimide, a non-stretchable material, by adjusting its thickness according to the curvature of the targeted anatomical area. From the electrochemical characterization of the devices, intrinsic thermal noise of platinum and gold electrodes is estimated to be 3-5 μV. Results show that electrode size and in vitro impedance do not influence the amplitude of the recorded SEPs. However, the use of a reference on-skull (a metal screw), as compared to reference on-array (a metal electrode surrounding the recording area), provides higher-amplitude SEPs. Additionally, the incorporation of a grounded metal shield in the thin-film devices limits crosstalk between tracks and does not compromise the recording capabilities of the arrays.

Ultrasensitive bioelectronic devices based on conducting polymers for electrophysiology studies

Conducting polymer electrodes based on poly(3,4-ethylenedioxythiophene doped with poly(styrenesulfonate) (PEDOT:PSS) are evaluated as transducers to record extracellular signals in cell populations. The performance of the polymer electrode is compared with a gold electrode. A small-signal impedance analysis shows that in the presence of an electrolyte, the polymer electrode establishes for frequencies below 100 Hz a higher capacitive electrical double layer at the electrode/electrolyte interface. Furthermore, the polymer/electrolyte interfacial resistance is several orders of magnitude lower than the resistance of the gold/electrolyte interface. The polymer low interfacial resistance minimizes the intrinsic thermal noise and increases the system sensitivity. The ultra-sensitivity of the polymer-based transducer system was demonstrated by recording the electrical activity of cancer cells of the nervous system.

Energy-Efficient Computing With Probabilistic Magnetic Bits—Performance Modeling and Comparison Against Probabilistic CMOS Logic

This paper focuses on the design, performance modeling, and evaluation of probabilistic logic using superparamagnetic nanomagnets for which there exists a strong interplay between deterministic dynamics and intrinsic thermal noise. The switching element in the spin domain is chosen as the giant spin-Hall effect (GSHE) device that operates based on the dipolar coupling phenomenon in a two-magnet system to achieve low-energy ( $\approx ~1.3$ fJ/b) and low-power ( $\approx ~0.5~\mu $ W) switching characteristics. The use of spin currents on the order of a few tens of mA/ $\mu $ m2 in the subcritical regime to operate the GSHE device yields nondeterministic switching behavior with probability of correctness less than 100%. Therefore, the proposed technique allows us to trade off circuit accuracy for tremendous reduction in energy and power dissipation. In this paper, we identify the required dimensions and material parameters of the read–write units to ensure robust magnetic dipolar coupling for reliable operation of the GSHE switch. Then, through Monte Carlo simulations, we evaluate the probabilistic switching behavior of the GSHE switch as a function of the input spin current amplitude and pulsewidth for various orientations of the magnetization vectors. The delay of the GSHE switch is quantified using a probability distribution function owing to the randomness imparted to the dynamics by intrinsic thermal noise of the nanomagnets. The relationship between the probability of correctness and the energy dissipation of the GSHE switch is quantified. The results are extended to evaluate the performance and circuit error rate of complex logic gates, such as NAND and NOR, constructed using the GSHE switch. It is shown that unlike the probabilistic CMOS (PCMOS) logic, the circuit error rate in the GSHE logic becomes a function of the input vector combination and the prior state of the switch. These nuances are captured in the compact model of the circuit error rate of multiple-input GSHE logic developed in this paper. The performance of the probabilistic GSHE logic is compared with that of PCMOS logic at the 14 nm technology node. Since the noise generation process in PCMOS logic has a limited bandwidth of tens of megahertz and consumes tens of microwatt power, the peripheral circuitry becomes prohibitive. By utilizing the inherent thermal stochasticity, nanomagnets provide a clear advantage to implement probabilistic computing platform targeted toward error-tolerant applications such as those from the image processing and machine learning domains.

Observation of fundamental thermal noise in optical fibers down to infrasonic frequencies

The intrinsic thermal noise in optical fibers represents the ultimate limit for fiber-based systems. However, at infrasonic frequencies, the spectral behavior of the intrinsic thermal noise is still unclear. In this letter, we present measurements of the fundamental thermal noise in optical fibers that are obtained using a balanced fiber Michelson interferometer. When an ultra-stable laser is used as the laser source and other noise sources are carefully controlled, the 1/f spectral density of the thermal noise is observed down to infrasonic frequencies, and the measured magnitude is consistent with the results of theoretical predictions at frequencies over the range from 0.2 Hz to 20 kHz. Moreover, as observed experimentally, the level of the 1/f thermal noise can be reduced by changing the coatings of the optical fibers. This therefore indicates one possible way to reduce thermal noise in optical fibers at low Fourier frequencies. Finally, the inconsistency between the experimental data and the existing theory for thermomechanical noise is discussed.

Temperature Dependence of IPRT and SPRT Self-Heating: Does Working with $${{\varvec{W}}}$$ W Really Help?

Platinum resistance thermometers (PRTs) are capable of providing reliable measurements at the millikelvin level, and are widely used in both industry and research applications. However, the intrinsic thermal noise associated with their resistance requires the use of a measurement current of typically around a milliampere to determine their resistance. Unfortunately, this same current also dissipates heat into the thermometer element, causing the well-known “self-heating” effect of typically a few millikelvins. Performing measurements in terms of the ratio to the resistance at the ice point provides some level of cancelation of this error around this temperature: If the thermal resistance between the sensor and environment were constant, this cancelation would work over a much wider temperature range. However, there is little evidence on the effectiveness of this strategy in practice. This paper reports on an extensive set of systematic measurements of the self-heating of six standard platinum resistance thermometers (SPRTs) and six industrial platinum resistance thermometers (IPRTs) of different designs, as a function of temperature, over the range from $$-190~^{\circ }\mathrm{C}$$ to $$420\,^{\circ }\mathrm{C}$$ , in a range of intercomparison baths and blocks. The measurements show that PRT self-heating varies from being almost constant with temperature to being nearly proportional to temperature. The assumption of a roughly temperature-independent thermal resistance is thus not justified in general. The results allow estimation of appropriate uncertainty terms for SPRT and IPRT self-heating for the two scenarios of “working in $$R$$ ” and “working in $$W$$ .”