Renaissance Man – The Transient Era
Although I had decided to enjoy retirement I had been approached earlier and agreed to return to NPL and consult part-time on the specific subjects for the Thermophysical Properties Group. These included reference materials, accreditation and certification requirements, international validation programmes, and distribution of information. I had similarly agreed to continue my association as a visiting consultant with Ulvac Riko in Japan on similar topics especially on non-contact transient methods and high temperature measurements on molten materials in space conditions. Both resulted in further consulting activities for an approximate ten-year period. However, I have since maintained continued interests in the subject through activities as a member of ASTM C-16 Thermal Insulation and E-37 Thermal Measurements Committees.
During the early years of this period at the NPL, through the Royal Society Foreign Guest Worker programme, I had the opportunity to work with several experts in relevant measurement topics. These included Ludovit Kubicar from Slovak Academy of Sciences in Bratislava, Kosta Maglic from Vinca Institute of Nuclear Science Belgrade, and Efim Litovsky originally from USSR Academy of Sciences. From these I obtained added benefit of their expertise in contact transient methodology, the flash method, and high temperature evaluation of porous materials, respectively. Similarly, in numerous visits to Japan I had the benefit of working with Akiharita Maesono and several of his colleagues at Sinku Riko (now Ulvac Riko) together with Professor Icihira Hatta at Nagoya University who all have considerable expertise in the development and application of both contact and non-contact transient methodology.
The renaissance period had indicated that the application often controlled the choice of the measurement method due to a sufficient amount and form of specimen being available. Thus, transient methods that require smaller amounts in disc or plate form for example 15 to 30 mm round or square and 1 to 10 mm thick method as required for the flash and contact transient techniques (CTS) were the obvious choice. Furthermore, cases arose where the specimen was now a film having thickness in the micron level on a substrate where thermal properties are thickness dependent. For these non-contact transient techniques were found more applicable.
In general, the contact transient methods (CTS) are based on modifications of the original absolute transient line-source (single, double, and crossed wire forms) technique. However, further variations of the transient plane source (TPS) with descriptive titles such and (hot-disc, -bridge, -strip and -ball) soon appeared. The only differences being the size form and shape of the energy source/temperature measurement sensor and the resultant analyses of different portions of the temperature/time response to the energy input.
The model for each technique is based on a solution with different assumptions, for each of the basic heat transfer equation in homogeneous semi-infinite bodies. Essentially a constant short heat pulse is generated symmetrically within a surrounding specimen to produce a temperature-time rise from which the slope is determined at a selected period(s) to derive one or more of the required thermal transport properties. They are often claimed individually not only to be all-encompassing having similar precision ranges, usually of the order of 3 to 5%, and applicable to all material types in the broad range covering the six orders of magnitude.
One commercial system had been acquired for investigation and evaluation at NPL as a secondary measurement tool. Measurements on numerous specimens of different solid materials that had been measured previously indicated that the precision and accuracy claims were, in general satisfied for a limited temperature range. However, when applied to materials, including non-solid thermal insulations having a thermal conductivity below 0,01 W/m.k, the performance range decreased below the 5% level. Since those early days, several improved models of the measurement system have been available with the same claims of precision and for a broad temperature range.
These methods are accepted as providing absolute results as for the case of the flash method, However, as will be discussed in a later blog on methods and apparatus a commercial comparative version so-called a modified transient plane source (MTPS) that requires the use of reference materials to obtain results of the same or improved accuracy was developed some 10 years ago. Although based now on asymmetric heat flow, lacking conformance to basic principles of physics, no independent verification of efficacy plus the use of fabricated reference materials of unknown provenance, it has become extremely popular in use and misuse for all material types having one stated specification for homogeneity.
The non-contact methods, such as the ac calorimeter are techniques based on working at frequencies, above the classical (quasi-static) limit, of thermal (or sound) wave transmission spectroscopy. In these, analyses of changes in the phase and amplitude of a modulated laser thermal wave transmitted through a plate-like sample are used to provide simultaneous determination of thermal diffusivity and specific heat capacity hence thermal conductivity. When the thermal contact conductance is taken into account the ac calorimeter method has been verified for materials having a thermal conductivity in the broad range 0.2–500 W/m.K and the added advantage of providing reliable and reproducible values for thin films of thickness less than 500 μm.
Although delayed by the 3000-mile move of family and home, the two-volume book Thermal Conductivity was finally published early in 1969.
In the preface to Volume 2, I had whimsically forecast “Until the ultimate dream is achieved of an all-purpose instrument for measuring the thermal conductivity of all materials over the full temperature range in less than a second with an accuracy better than half of a percent we must rely on the experience gained with existing reliable methods”.
Thus, on my return to the NPL Thermal Properties Group it seemed most appropriate to be involved immediately with developing new reference materials applicable for existing applications. At the time separate programmes were underway in developing high temperature and high conductivity (>10 W/m.K) nickel and iron alloys and a dense low conductivity (< 0.25 W/m.K) polymethylmethacrylate for a 100K degree range above ambient. Several workers especially in Japan were investigating different forms of transient methods to measure thermal properties of thin films of diamond, carbon and silicon at the micron thickness level. The broad ranges and scatter of values obtained by these different methods were reduced only when Tetsui Baba and colleagues at the Advanced Industrial Science and Technology Institute (AIST) developed a stable silicon dioxide thin film reference artefact.
In this context, mention of the ranges of thermal conductivity and temperature is relevant and most important since it must be realised that these are a total of six and three orders of magnitude respectively. Unfortunately, at the present time the total number of such artefacts is limited to less than twenty, most are pure metals or iron and nickel alloys that cover a broad temperature range while many having known with values only for a narrow range of use. This is not sufficient for such an important requirement and there is an urgent need for several more especially in the range 0.1 to 5 W/m.K for use at higher ranges of temperatures and more immediately at least one thermal insulation for the temperature range to 800K.
The reason for this lack of artefacts is cost and time of development. As an example, in the late nineties the NPL group obtained a European Union funded contract to manage a programme to review, select and develop a reference material for high temperatures above 1000K. Following a relatively short review Pyroceram 9606 became the obvious candidate Overall, a total of twelve organisations from government, commerce and academia were involved from the first day at a total high six-figure cost and taking a total time approaching four years in the following:
- Assessing, and obtaining the estimated “lot” for testing and a 10+ year sale stock;
- Characterization process measurements establishing stability and reproducibility
- Repeat measurements by three or more sources of different thermal conductivity absolute methods
- Support thermal diffusivity, specific heat and thermal expansion measurements
- Individual organization and evaluation of the results to the final
- Evaluation and certification by Institute for Reference Measurement and Materials
As mentioned in the second Blog, during the renaissance years the significant expansion of national and international trade required a “leveling of the playing field” in many fields. These stimulated the need, especially in areas related to energy conservation for certification of product performance and accreditation of in-house and external measurement organizations. First, by the end of the century in both Europe and the USA, stringent legal thermal performance specifications of a maximum uncertainty 5% or better had been applied to insulation products for use in buildings.
Essentially these required the need for measurement to be undertaken using apparatus that guaranteed an accuracy of better than 5% for the temperature 250 to 350K. In Europe this was extended in 2012 to insulation products for what is termed industrial applications in the temperature range 100 to 800K. As NPL had appropriate heat flow meter and a two guarded hotplate apparatus for use associated with the building products and a large one for the higher temperatures, these were in constant use by my colleague David Salmon and I on projects to support attainment of these legal requirements.
I referred to the “revolution” in measurements of performance of thermal insulations for building applications due to the introduction and subsequent worldwide use of the heat flow meter method, particularly by industry for QE and QA requirements. Essentially the salient feature of this “simple” method is that the claimed 4 to 5% precision level is wholly dependent on the reference materials having precision used a level of 2% or better as claimed by NIST for SRM 1450(a-d) in the USA and by LNE for CRM 440 in France.
In this context, in 1997 at an ASTM C-16 Conference on Guarded Hot Plate and Heat Flow Meter Technology that I and Cliff Shirtliffe from NRC Canada organized I suggested and offered to organize an inter-laboratory comparison of the national guarded hot plates of Canada, France, Japan, UK and USA. This was agreed that measurements were to be made on the same specimens over the approximate 280 to 380K range temperature range of each plate. NIST would provide two specimens of SRM 1450c, a 25mm thick bonded fibrous glass board and SRM 1453, a 13 mm thick expanded polystyrene board.
The results were surprisingly positive illustrating that despite differences in apparatus size, form and effects of packaging and distance travelled, a precision/accuracy level close to the +/-1% with repeatability of better than 1% were attained for all apparatus. These results were sufficient evidence to support the 3 to 4% claim for the general use of the method providing these references values of +/-2% or better were used. As a result we organised similar comparisons on the five UK testing organisations that were accredited for use of the guarded hot plate method and the seven using the heat flow meter. The results, now confirmed regularly by individual spot-checks when being accredited, did indicate that all organizations attained and maintained levels that were within the required values.
Fortuitously, SRM 1450b, a 25mm thick high density fibrous glass in the USA and CRM 440, a 37mm thick selected lower density fine glass fibre blanket in Europe have been available at the time for the important calibration role for 25 to 50 mm thick specimens for the range 270K-350K. Although these materials exhibit significant anisotropy they have highly stable properties in the heat-flow perpendicular to fibre direction. It is now necessary to have even thicker calibration specimens for products above 50 mm thick. NIST in the USA and NPL in Europe each have apparatus (0.9m to 0.6m) and able to provide 150mm thick individual Transfer Standards TS similar fibrous specimens having unique measured thermal resistance values at the 1-2% precision level for the calibration of such large apparatus.
It is unfortunate that these required performance levels cannot be claimed for temperatures below 250K and more certainly above 350K levels due, essentially to the lack of available reference artefacts. The reason for this factor is that the accuracy and precision levels for the guarded hot plate at these temperatures remain highly questionable compared with those described above. Future Blogs on methods and reference materials will discuss the many conflicting issues involved and means of their resolution.
An interesting and probably forgotten relevant fact is that prior to 1970 the most frequently used ASTM Standard C-177 for the method consisted of two parts, described as low and high temperature forms of apparatus for temperatures below and above 400K. The present all-encompassing revised standard for the whole temperature range was published in 1970. While the original apparatus were similar significant details regarding the form, the materials of construction and change in operation were included in the new version. An ASTM C-16 inter-laboratory comparison involving five pre-1970 hot plates had indicated that the maximum difference in the results at 900k was some 20% and above 30% at 1200k.
Between 1980 and 2018, further reviews of the method and comparison of measurements have been undertaken by members of the ASTM C-16 Committee (twice), the ISO TC-163, and two European groups the last of which included all hot plates of national measurements organizations or their equivalents. During the approximate span of 50 years and following each of the reviews, the differences have decreased slow to 5% and 12 % respectively. While just acceptable for current requirements there is an obvious need both for improvement in this value and assurance of consistency and reproducibility.
Early in the transient era it appeared from claims of performance of the transients that major steps could be taken towards fulfilling the desired quantitative levels for these newer methods. However, uncertainty and disappointment quickly arose due to apparent significant differences (>8% to 10% with extremes above 25%) obtained not only between values using a transient method and an absolute or accepted method but also between two transient methods. This may be explained by the continuing development and particularly the use, by inexperienced workers of commercial fully automated forms that use rote pre-determined processes in measurement irrespective of the type and form of specimen. The subject is complex, many factors including for example, specimen size, form, homogeneity, heterogeneity radiative and convective heat transfer have to be considered in the choice and use of the most appropriate method.
Certainly, this is one area where a comprehensive totally independent inter-laboratory measurements programme at international level should be undertaken without delay. While the study should involve each technique and the currently available SRMs, there should be agreement, with the cooperation of one or more national measurement laboratories, to select and measure a small number (say 4 to 5) of recommended additional reference materials.
There is no doubt that various arising global factors and their resulting needs have been responsible for the significant changes and improvements in this complex subject and the corresponding advances that have occurred during these three eras. These have resulted not only in methodology but more importantly in the attitude to and understanding of the different mechanisms of transfer of heat in materials. Thermal conductivity is no longer assumed to be viewed as a “simple” esoteric, easily measured, property for inclusion in textbooks and collections of data of typical values for different materials. It has become a definitive property to be considered in relation the design, development and economic operation of systems ranging from a building to a nuclear reactor to a computer component. However, its attainment to idealist limits remains o lucrare în curs de desfășurare.
Modified Transient Plane Source (MTPS), Source: Wikipedia