Search

This paper considers the influence, in terms of temperature and time, of the following three factors on the phase transition of high-temperature fixed points: the temperature gradient of the furnace along the fixed-point cell, the thermal inertia of the furnace, and the thermal condition of the cell itself. Melting and freezing experiments have been performed for the eutectic Co–C in a three-zone furnace, of which the temperature gradient along the cell could be varied from positive to negative. To help the interpretation of these experiments, a realistic 2D finite-element model has been implemented, simulating the three influence factors. The model allowed us to understand the features observed in the melts and associated freezes, for the different furnace conditions, not only qualitatively but also quantitatively. This research identifies the essential requirements to facilitate the optimum configuration of the furnace and cell design.

Radiometric techniques for temperature measurements are indispensable in industrial applications, particularly when the use of contact thermometers is hard or impossible to realize. The principles and realizations of some new and extended radiometric techniques for measuring the emissivity and temperature of an object are presented. Using the described techniques, the emissivity and temperature of an Inconel 600 sample at high temperatures in laboratory conditions were determined. The validation of the temperature measurement of the same sample in a simulated industrial condition is also presented.

A multi-partner project to determine the thermodynamic temperatures of a selected set of high-temperature fixed points based on metal-carbon eutectics is underway as a working group activity within the Comite International des Poids et Mesures. The investigation focuses on four fixed-point types, namely, the three metal-carbon eutectic points of Re-C $$(2474\,^{\circ }\hbox {C})$$ , Pt-C $$(1738\,^{\circ }\hbox {C})$$ , and Co-C $$(1324\,^{\circ }\hbox {C})$$ , and the Cu point $$(1084.62\,^{\circ }\hbox {C})$$ . This paper describes the construction, pre-evaluation, and screening stage of the cells prior to their thermodynamic temperature determinations. The construction of the HTFP cells was undertaken by nine national metrology institutes (NMIs) according to instructions laid out in a pre-agreed protocol that ensures production of best quality cells. Four NMIs conducted the evaluation, each for a certain fixed-point type, and screened out cells that did not meet pre-determined selection criteria while assuring sufficient variety in the sources of the cells in the final selected sets. In autumn 2012, the selected cells were successfully passed on to the final stage of the project, the thermodynamic temperature measurement, and assignment.

Hybrid cells with double carbon/carbon sheets are used at the Istituto Nazionale di Ricerca Metrologica (INRIM) for the realization of both pure metal fixed points and high-temperature metal–carbon eutectic points. Cells for the Cu and Co–C fixed points have been prepared to be used in the high-temperature fixed-point project of the Comite Consultatif de Thermometrie. The results of the evaluation processes were not completely satisfactory for the INRIM cells because of their low transition temperatures with respect to the best cells, and of a rather large melting range for the Co–C cell. A new design of the cells was devised, and considerable improvements were achieved with respect to the transition temperature, and the plateau shape and duration. As for the Cu point, the duration of the freezing plateaux increased by more than 50 % and the freezing temperature increased by 18 mK. As for the Co–C point, the melting temperature, expressed in terms of the point of inflection of the melting curve, increased by about 70 mK. The melting range of the plateaux, expressed as a difference $$(T_{\mathrm{liquidus}} - T_{\mathrm{solid}})$$ was reduced from about 180 mK to about 130 mK, with melting times $$(t_{\mathrm{liquidus}} - t_{\mathrm{solid}})$$ increased by about 50 %, as a consequence of an improvement of flatness and run-off of the plateaux.

In principle, multiwavelength radiation thermometry allows one to correctly measure the temperature of surfaces of unknown and varying surface emissivity. Unfortunately, none of the practical realizations proposed in the past proved to be sufficiently reliable because of a strong influence of the errors arising from incorrect modeling of the emissivity and of the limited number of operating wavelengths. The use of array detectors allows a high degree of flexibility both in terms of number and spectral position of the working wavelength bands. In the case of applications at high temperatures, i.e., near 2000 $$^{\circ }$$ C or above, an analysis of the theoretical measuring principles of multiwavelength thermometry, suggests the opportunity of investigating the possible advantages in extending the operating wavelengths toward the ultraviolet region. To this purpose, a simulation program was developed which allows investigation of the effect of different influencing parameters. This paper presents a brief theoretical introduction and practical analysis of the method. The best choices are derived in terms of the different influencing parameters and data relative to the simulation of both real materials and fictitious emissivity curves and have been studied and analyzed with different emissivity models to check the robustness of the method.

Over the medium temperature range (from $$156\,^{\circ }\mathrm{C}$$ to $$1000\,^{\circ }\mathrm{C}$$ ), radiation thermometry is usually established within a national metrology institute (NMI) by means of variable temperature blackbody radiation sources, whose temperature is determined using a platinum resistance thermometer or thermocouple, calibrated in terms of the International Temperature Scale of 1990 (ITS-90), positioned in close proximity to the back radiating surface of the blackbody. It is also reasonably common to establish a scale using a suitable radiation thermometer, such as an indium gallium arsenide (InGaAs) detector-based narrow band radiation thermometer, calibrated using a number of fixed-point blackbody sources from the indium (In) to silver (Ag) (or copper (Cu)) points, with the calibration results fitted using a parameterized Planckian interpolation function. During 2007 and 2008, two InGaAs-based radiation thermometers were circulated around seven NMIs within the European Association of National Metrology Institutes (EURAMET) region in order to undertake a comparison of parameters necessary for radiation thermometry over the medium temperature range. Measurements were made of the size-of-source effect and gain (range) ratios of the two thermometers along with an assessment of the effect of changes in the ambient temperature and humidity on the thermometer output. The thermometers were also calibrated using fixed-point and/or variable temperature blackbody sources at each institute. A brief overview of the results obtained by this project is presented in this paper.

Two hybrid cells for realizing the Co–C and Pd–C fixed points and constructed at Istituto Nazionale di Ricerca Metrologica (INRIM) were used for an evaluation of facilities and procedures adopted by INRIM and National Metrology Institute of Singapore (NMC) for the realization of the solid–liquid phase transitions of high-temperature fixed points and for determining their transition temperatures. Four different furnaces were used for the investigations, i.e., two single-zone furnaces, one of them of the direct-heating type, and two identical three-zone furnaces. The transition temperatures \(T_{90}\) were measured at both institutes by adopting different procedures for realizing the radiation scales, i.e., at INRIM a scheme based on the extrapolation of fixed-point interpolated scales and an International Temperature Scale of 1990 (ITS-90) approach at NMC. The point of inflection (POI) of the melting curves was determined and assumed as a practical representation of the melting temperature. Different methods for deriving the POI were used, and differences as large as some hundredths of a kelvin were found with the different approaches. The POIs of the different melting curves were analyzed with respect to the different possible operative conditions with the aim of deriving reproducibility figures to improve the estimated uncertainty. As regard to the institutes inter-comparison, differences of 0.13 K and 0.29 K were found between INRIM and NMC determinations at the Co–C and Pd–C points, respectively. Such differences are compatible with the combined standard uncertainties of the comparison, which are estimated to be 0.33 K and 0.36 K at the Co–C and Pd–C points, respectively.

INRIM and LNE-Cnam agreed to undertake a collaboration aimed to verify, through the use of metal–carbon eutectic fixed-point cells, methods and facilities used for defining the transition temperature of eutectic fixed points and manufacturing procedures of cells. To this purpose and as a first step of the cooperation, a Co–C cell manufactured at LNE-Cnam was measured at INRIM and compared with a local cell. The two cells were of different designs: the INRIM cell of 10 cm3 inner volume and the LNE-Cnam one of 3.9 cm3. The external dimensions of the two cells were noticeably different, namely, 40 mm in length and 24 mm in diameter for the LNE-Cnam cell 3Co4 and 110 mm in length and 42 mm in diameter for the INRIM cell. Consequently, the investigation of the effect of temperature distributions in the heating furnace was undertaken by implementing the cells inside single-zone and three-zone furnaces. The transition temperature of the cell was determined at the two institutes making use of different techniques: at INRIM radiation scales at 900 nm, 950 nm, and 1.6 μm were realized from In to Cu and then used to define T90(Co–C) by extrapolation. At LNE-Cnam, a radiance comparator based on a grating monochromator was used for the extrapolation from the Cu fixed point. This paper presents a comparative description of the cells and the manufacturing methods and the results in terms of equivalence between the two cells and melting temperatures determined at INRIM and LNE-Cnam.

Three different Pd-C eutectic fixed-point cells were prepared and investigated at INRIM. Several tens of phase transition runs were carried out and recorded with both a Si-based radiation thermometer at 950 nm and a precision InGaAs-based thermometer at 1.6 μm. Two of the cells were of the same design with an inner volume of 12 cm3. The third one was smaller with a useful inner volume of 3.6 cm3. The three cells were filled with palladium powder 4N5 or 4N8 pure and graphite powder 6N pure. The repeatability and stability of the inflection point were investigated over a period of 1 year. The noticeably different external dimensions of the two cells, namely, 110 mm and 40 mm in length, allowed the influence of the longitudinal temperature distribution to be investigated. For this purpose, two different furnaces, a single-zone with SiC heaters and a three-zone with MoSi2 heaters, were used. Different operative conditions, namely, temperature steps, melting rate, longitudinal temperature distributions, and position of cells within the furnace, were tested to investigate the reproducibility of the cells. Effects on the duration and shape of the plateaux were also studied. This article gives details of the measurement setup and analyses of the melting plateaux obtained with the different conditions.

A comparison of a Pt/Pd calibration above the Ag point between the INRIM and NMC was arranged with the aims of evaluating measurement systems and exploiting the potential of the Pt/Pd thermocouples. Two commercial Pt/Pd thermocouples were used as transfer thermometers. A calibration method using a blackbody cavity as a transfer source and a radiation thermometer as a reference thermometer was adopted in both institutes. The T 90 carried by the radiation thermometers is established by an extrapolation technique for INRIM and by scale realization according to ITS-90 definition for NMC and, therefore, this exercise is also a useful comparison of different approaches to disseminate T 90 above the Ag point. The comparison results are presented and analyzed.

Since the value T90(Au) was fixed on the ITS-90, some determinations of the thermodynamic temperature of the gold point have been performed which form, with other renormalized results of previous measurements by radiation thermometry, the basis for the current best estimates of (T − T90)Au = 39.9 mK as elaborated by the CCT-WG4. Such a value, even if consistent with the behavior of T − T90 differences at lower temperatures, is quite influenced by the low values of TAu as determined with few radiometric measurements. At INRIM, an independent indirect determination of the thermodynamic temperature of gold was performed by means of a radiation thermometry approach. A fixed-point technique was used to realize approximated thermodynamic scales from the Zn point up to the Cu point. A Si-based standard radiation thermometer working at 900 nm and 950 nm was used. The low uncertainty presently associated to the thermodynamic temperature of fixed points and the accuracy of INRIM realizations, allowed scales with an uncertainty lower than 0.03 K in terms of the thermodynamic temperature to be realized. A fixed-point cell filled with gold, 99.999 % in purity, was measured, and its freezing temperature was determined by both interpolation and extrapolation. An average TAu = 1337.395 K was found with a combined standard uncertainty of 23 mK. Such a value is 25 mK higher than the presently available value as derived by the CCT-WG4 value of (T − T90)Au = 39.9 mK.

The calibration of platinum-based thermocouples from 420 °C to 1,100 ˚C is currently carried out at INRIM making use of two different apparatus: for temperatures below 930 ˚C, a potassium gas-controlled heat pipe (GCHP) is used, whereas a metal-block furnace is adopted for higher temperatures. The standard uncertainty of the reference temperature obtained in the lower temperature range is almost one order of magnitude better than in the higher temperature range. A sealed copper cell was investigated to see if it could be used to calibrate thermocouples above 930 ˚C with a lower uncertainty than our current procedures allowed. The cell was characterized with Type S and Pt/Pd thermocouples and with an HTPRT. The freezing plateaux were flat within 0.01 ˚C and lasted up to 1 h with a repeatability of 0.02 ˚C. The temperature of the cell was determined with a standard uncertainty of 0.04 ˚C. Hence, the copper cell was found to be superior to the comparator furnace for the calibration of platinum-based thermocouples because of the significant decrease in the uncertainty that it provides. An analysis was also carried out on the calibration of Pt/Pd thermocouples, and it was found that the combined use of the potassium GCHP and the Cu fixed-point cell is adequate to exploit the potential of these sensors in the range from 420 °C to 1,084 °C. A comparison with a fixed-point calibration was also made which gave rise to agreement within 0.07 ˚C between the two approaches.

The high-temperature extension of the fixed-point technique for primary calibration of precision infrared (IR) thermometers was investigated both through mathematical simulations and laboratory investigations. Simulations were performed with Co–C (1,324°C) and Pd–C (1, 492°C) eutectic fixed points, and a precision IR thermometer was calibrated from the In point (156.5985°C) up to the Co–C point. Mathematical simulations suggested the possibility of directly deriving the transition temperature of the Co–C and Pd–C points by extrapolating the calibration derived from fixed-point measurements from In to the Cu point. Both temperatures, as a result of the low uncertainty associated with the In–Cu calibration and the high number of fixed points involved in the calibration process, can be derived with an uncertainty of 0.11°C for Co–C and 0.18°C for Pd–C. A transition temperature of 1,324.3°C for Co–C was determined from the experimental verification, a value higher than, but compatible with, the one proposed by the thermometry community for inclusion as a secondary reference point for ITS-90 dissemination, i.e., 1,324.0°C.

Below the freezing point of silver, radiation thermometers are generally calibrated by implementing the multi-point interpolation method using blackbody measurements at three or more calibration points, rather than the ITS-90 extrapolation technique. The interpolation method eliminates the need to measure the spectral responsivity and provides greater accuracy at the longer wavelengths required below the silver point. This article identifies all the sources of uncertainty associated with the interpolation method, in particular, those related to the reference blackbody temperatures (either variable-temperature or fixed-point blackbodies) and to the measured thermometer signals at these points. Estimates are given of the ‘normal’ and ‘best’ uncertainties currently achievable. A model of the thermometer response is used to propagate all the uncertainties at the reference points and provide a total uncertainty at any temperature within the calibration range. The multi-point method has the effect of constraining the total uncertainty over this range, unlike the ITS-90 technique for which the uncertainties propagate as T 2. This article is a joint effort of the working group on radiation thermometry of the Consultative Committee for Thermometry (CCT), summarizing the knowledge and experience of all experts in this field.

At INRIM, different Co–C fixed-point cells have been constructed and investigated. Two cells of different design and volume and filled with highly pure cobalt (99.998%) were used to extend the fixed-point calibration of five Pt/Pd thermocouples that had been previously calibrated at the triple point of water and at the fixed points of In, Sn, Zn, Al, and Ag. The calibration at the Cu point was also added during this exercise. Because a previous calibration from 962 °C up to 1,500°C against the local standard radiation thermometer was available, a comparison was possible with the Co–C fixed-point calibration. Agreement within 0.10 °C was found when the value of 1,324.0 °C, the same value proposed for the Co–C point to be included as a secondary reference point of the ITS-90, was assumed.

The behavior of the Co–C eutectic fixed point was investigated at INRIM. Several cells of different design and volume, and filled with cobalt of different purity were constructed and investigated with both Pt/Pd thermocouples and radiation thermometers. The melting behavior was investigated with respect to the melting rate, the pre-freezing rate, and the annealing time. The melting temperatures, as defined, were not significantly affected by the different testing conditions, even if the shape and duration of the plateaux were influenced. Several tens of melt and freeze cycles were performed with the different cells. The spread in the results for all of the different conditions was very limited in extent, giving rise to a standard deviation of less than 0.04 °C; a repeatability of better than 0.02 °C was found with both Pt/Pd thermocouples and radiation thermometers. The results of our measurements are encouraging and confirm the suitability of Co–C as a reference point for the high-temperature range in a possible future temperature scale. Investigations of long-term stability remain ongoing.