"Johnson noise thermometry (JNT)" is a technique for measuring temperature that is based on the fundamental properties of thermal fluctuations in conductors. A Johnson noise thermometer does not require calibration and is not affected by the condition of the sensor material. Therefore, it is well-suited for long-term temperature measurements in harsh environments, such as nuclear reactor coolant circuits, nuclear waste management, and storage. It is also useful in situations where low drift is necessary, such as high-temperature reference standards in metrology or the annealing of single-crystal turbine blades in the aerospace industry.
Currently, all thermometers used, including thermocouples, platinum resistance thermometers, and thermocouples, are considered secondary thermometers and are therefore susceptible to drift. These thermometers do not directly measure temperature, but instead measure a property of the sensor, such as resistance or EMF. The relationship between this property and temperature is determined through a calibration process. To convert the measurement back to temperature, this relationship is used, assuming it has not changed. However, in harsh environments, such as those with sensor contamination, physical structure changes, or transmutation in nuclear environments, the relationship can change over time, leading to "drift" in the reported temperature.
Primary thermometers, on the other hand, are fundamentally different. With a primary thermometer, all necessary sensor properties are measured and then used in a fundamental physical law to calculate the true thermodynamic temperature. Since the measurement of temperature is based on a fundamental physical law, primary thermometers do not drift. While electronics used to measure the required properties may drift, modern electronics are sufficient to ensure that such drift is insignificant. In practical thermometers today, it is usually the sensor, rather than the electronics, that leads to drift.
Johnson noise, also known as thermal noise, Johnson-Nyquist noise, or white noise, is a type of electronic noise that is generated by the thermal agitation of charge carriers, typically electrons, within an electrical conductor. This noise is present even when no voltage is applied to the conductor. The fluctuation-dissipation theorem is the statistical physical derivation that explains the generic nature of this noise.
In an ideal resistor, Johnson noise is considered white, which means that the power spectral density remains constant throughout the frequency spectrum, except at extremely high frequencies. Since Johnson noise is the result of many independent charge carrier movements, the central limit theorem states that the resulting noise voltage will have a normal or Gaussian distribution.
The most significant feature of Johnson noise is that its power is directly proportional to absolute temperature. This means that electrical measurements can be used to derive the absolute temperature without the need for material properties or calibration requirements. As a result, this technique is not subject to sensor drift.
To create a thermometer that is not subject to drift, the Johnson noise in a resistive sensor can be measured along with the sensor resistance and measurement bandwidth. Using the Johnson-Nyquist equation, it is then possible to determine the true thermodynamic temperature of the sensor, independent of the sensor's state:
The Johnson noise signal is incredibly small, comparable to the electrical noise found in the best low-noise amplifiers. This makes it extremely challenging to measure with the necessary precision. It is also susceptible to contamination by electrical noise in the environment. Despite many attempts to create a practical Johnson noise thermometer since the advent of modern electronics in the late 1950s, no commercial product has been developed due to the difficulties associated with low signal levels and signal contamination by external noise.
Metrosol, in collaboration with the National Physical Laboratory (NPL), is currently developing the world's first practical Johnson Noise Thermometer. This work began in 2014 with a scoping and feasibility project, which led to the creation of a completely new technique for measuring Johnson noise. The new technique was patented, and a demonstrator/proof-of-concept prototype was developed. The demonstrator operated over the range of -20 to 120°C, used several commercial instruments to make measurements, and weighed 35kg in a 70L volume.
From 2017 to 2020, we developed all the necessary core technology, including electronics, software, mechanics/enclosure, and a high-temperature probe, to create a practical thermometer capable of operating up to 1,000°C in a commercial format weighing 1kg and having a volume of 1L.
One of the most significant accomplishments during this phase was that when the prototype was tested for electromagnetic compatibility (EMC), it was found to be entirely resistant to external electromagnetic radiation up to the heavy industrial limits specified in EN61000-4-3.
In 2021, we have been analyzing all the sub-systems of the Johnson Noise Thermometer to optimize the design and create a comprehensive uncertainty budget for the thermometer. By the end of the year, we will have all the necessary technology to produce a commercial Johnson Noise Thermometer. We will then begin working with potential adopters to develop and produce the first commercial product using this new technology.