Reference Materials for Thermal Conductivity Measurements: European Situation

Abstract

A reference material (RM), as defined by the International Vocabulary of Metrology (VIM 2012), must be homogeneous, stable, and suitable for use in measurements. Certified reference materials (CRMs) are RMs with documented property values, uncertainties, and traceability. ISO 17034:2018 outlines the requirements for RM producers, ensuring that CRMs meet standards for stability, uniformity, and reproducibility. In Europe, CE marking, from French “conformité Européenne”, which means European conformity, has been mandatory for thermal insulation products since 2002, ensuring their thermal performance is verified by accredited laboratories using RMs like IRMM440 and ERM FC440. Annually, European manufacturers produce over 200 million cubic meters of thermal insulation, necessitating thousands of thermal conductivity measurements daily to maintain CE marking compliance. Key characteristics of Reference Materials include long-term stability, thermal conductivity within specified ranges, and minimal dependence on density, thickness, and applied load. Sample thickness must conform to apparatus specifications, and homogeneity must be quantified. Reference Materials must also have appropriate dimensions, surface smoothness, and manufacturability. The Joint Research Centre (JRC) Geel has produced two Reference Materials, IRMM 440 and ERM FC 440, with specific characteristics to meet these requirements. Both are glass wool fibers with low thermal conductivity and specific density and thickness. The qualification of RMs involves inter-laboratory comparisons to ensure the accuracy and traceability of thermal conductivity measurements. The European market’s organization, including the use of Reference Materials and CE marking, has significantly improved measurement consistency and product quality. This system has led to lower uncertainties in thermal conductivity measurements compared to North America, highlighting the impact of standardized RMs on industry practices. Future needs include developing RMs with lower conductivity and increased thickness to accommodate market trends towards super insulation materials and bio-based components, enhancing energy performance calculations for buildings. This paper will present the process of defining a reference material and how it affects the uncertainty level of the calculation of building energy performance. This level depends on the characteristics of the materials used, their implementation, and external factors, such as the weather, as well as the reference material used for calibration of all European thermal conductivity measurement devices.

1. Introduction

In the International Vocabulary of Metrology—Basic and General Concepts and Associated Terms (VIM 2012), a reference material (RM) is defined as a material that is sufficiently homogeneous and stable with reference to specified properties, and which has been established to be fit for its intended use in measurement or examination of nominal properties. In the same document, a certified reference material (CRM) is a reference material that is accompanied by documentation issued by an authoritative body, providing one or more specified property values with associated uncertainties and traceability, using valid procedures.

The standard ISO 17034:2018 [1] outlines general requirements for the competence of reference material producers. It summarizes requirements regarding the competence of personnel, production, sample handling and storage, and aspects directly related to material preparation. Certified reference materials that meet ISO 17034:2018 guidelines for long-term property stability, uniformity, and reproducibility should allow comparison of different apparatus and techniques.

A wide variety of methods exist for measuring the thermal conductivity of insulation materials, including the guarded hot plate (GHP) and Heat Flow Meter (HFM) methods. The review by Numan Yüksel (2016) [2] provides a comprehensive overview of these methods, highlighting their application in determining the thermal performance of insulation products. Similarly, Hameed’s (2024) [3] review underscores the significance of selecting appropriate insulation materials to reduce thermal loads in buildings, emphasizing the critical role of accurate thermal conductivity measurements.

In Europe, the CE marking has been mandatory since 2002 for thermal insulation products for buildings and since 2012 for thermal insulation products for industrial installations. This means that the thermal performance of products is verified by an accredited and notified laboratory. The reference material to be used in the process of calibration of the Heat Flow Meters (HFMs) or guarded hot plates (GHPs) apparatuses [2] shall be the materials cited in the EN 13172 standard [4], i.e., IRMM440 in the current version and ERM FC440 for in the revised version which will be published at the end of 2023.

Each year, European manufacturers produce more than 200 million cubic meters of thermal insulation products. The most common types of materials used for insulation of buildings include mineral wool (e.g., stone or glass wool) and plastic foams (e.g., polyurethane, polystyrene, etc.) (

Table 1. List of most common thermal insulation products having a harmonised standard (hEN) or a European technique assessment (ETA).

Inorganic Mineral	Organic Fossil Fuel Derived	ORGANIC PLANT
Stone wool Glass wool (MW)	Polyurethane (PU)	Wood fiber (WF)
Cellular glass (CG)	Expanded polystyrene (EPS)	Cork (ICB)
Expanded Perlite (PEB)	Extruded Polystyrene (XPS)	Wood wool (WW)
	Phenolic foam (PF)	Cellulosic product (CI)

and

Table 2. List of most common thermal insulation products without a harmonised standard (hEN) or European technique assessment ETA.

Innovative Material	Organic Plant
Aerogels	Cotton
Vacuum insulation panel (VIP)	Hemp
Phase change	Straw
Reflective insulation product	Sheep wool

). Each day, thousands of thermal conductivity measurements are made by the manufacturers. These values are the basis for controlling production to be in line with the CE declaration.

The European reference material is the basis for the thermal conductivity level for all European laboratories and manufacturers, and it has a big impact on the production of each thermal insulation plant in Europe, and ultimately on the energy performance of buildings. Moreover, the reference material is used by accredited laboratories for calibration, internal quality control, performance checks, and also in participation in interlaboratory tests. Therefore, the choice and methodology for defining a new reference material is critical and should be performed carefully.

Zarr (2008) [5] and Porzuczek (2024) [6] have also discussed the importance of developing and maintaining reference materials for thermal insulation. Zarr highlights the status and evolution of NIST thermal insulation reference materials, noting their historical development and ongoing relevance (Zarr, 2008) [5]. Porzuczek’s study provides a comparative analysis of selected insulating materials for industrial piping, further reinforcing the need for accurate and reliable reference materials in various applications (Porzuczek, 2024) [6].

In daily life, the quality of a reference material is important for the production cost. If the lambda meter has an error of 3%, or 0.001 W/m·K, the effect on the production cost can reach 20% due to the change in the density of the product. On the energy performance of building (EPBC), the effect is directly 3% of the energy consumption.

This paper will describe the process used in Europe for the last two reference materials, the consequence on the European market of insulation, and the future needs of the European industry.

2. Characteristics of a Reference Material

2.1. Expected Thermal Conductivity

From the VIM and the ISO 17034, the reference material should at least meet the following criteria:

• Long-term stability. The aim of this characteristic is to guarantee a stable value of thermal conductivity of the specimen for several years. The long-term stability of a candidate reference material can be evaluated using an accelerated process such as an increase in temperature or moisture cycling. The aging method varies depending on the type of material selected. It can also be based on common knowledge from measurements performed over several years. This condition does not allow for the use of materials that age over time, such as cellular plastics with an expansion gas whose conductivity is lower than air or vacuum insulation products. The most common of these products are polyurethane foam [7], extruded polystyrene [8], and phenolic foam [9]. They are produced with an expansion gas, which disappears from the product with time. The evolution of their thermal conductivity can be above 0.006 W/(m·K), which is not acceptable for a reference material. For vacuum insulation products, the evolution of thermal conduct is due to the air permeation through the envelope. Like cellular plastic, the increase in thermal conductivity can be important.
• Thermal conductivity should be in the range of thermal insulation materials sold in Europe or produced in the plant, i.e., in general between 0.02 and 0.04 W/(m·K) at 10 °C mean temperature. As written above, products having a thermal conductivity below air are products with a blowing agent. These products age with time and then can not be used as reference material. Today, the only possible reference material can be between 0.3 and 0.4 W/(m·K). A product made of aerogel composed of a mix between standard thermal insulation material and aerogel could be a possible candidate, but such a product has not been evaluated yet.
• There is a low dependence of thermal conductivity on density, thickness, and applied load in the device. This will reduce the uncertainty of thermal conductivity stemming from production parameters.

2.2. Effect of Thickness

The effect of thickness occurs at two levels, apparatus performance and infrared radiation going through the material.

2.2.1. Minimum and Maximum Thickness Due to Apparatus Performance

The minimum sample size is the smallest size of the specimen that should be used in a lambda-meter apparatus, a heat flow meter of a guarded hot plate. The minimum thickness is linked with the gap size of the apparatus (distance between the metering area and the guard) and should be at least ten times the width of the gap. The minimum thickness is also linked by thermal contact resistances and shall not induce an uncertainty greater than 0.5% of the thermal conductivity [10]. The minimum thickness is stated in

Table 3. Minimum and maximum allowed specimen thickness [10].

Overall Size (mm)	Metering Section (mm)	Guard Width (mm)	Maximum Thickness [Edge Limit] (mm)	Minimum Thickness [Gap Limit] (mm)
200	100	50	30	12.5
300	200	50	35	25.0
300	150	75	45	18.8
400	200	100	60	25.0
400	100	150	80	12.5
500	300	100	65	37.5
500	250	125	75	31.3
500	200	150	85	25.0
600	300	150	90	37.5
800	500	150	100	62.5
800	400	200	120	50.0
1000	500	250	150	62.5

and comes from EN 12667:2001 [10].

The maximum thickness of the samples is stated in the EN 12667:2001 standard and is linked with the edge losses of heat flow between the specimen and the environment (see Table 3). If the thickness is too high, a large part of the heat flow coming from the hot plate will dissipate into the environment. Also, above this value, the heat flow will not be strictly perpendicular to the hot plate.

Considering these factors, the thickness of a product sold in Europe and in accordance with the heat flow meter specifications used in plants for factory product control should be between 20 mm and 150 mm, depending on the size of the apparatus (see Table 3 from EN 12667:2001 standard).

2.2.2. Effect of Thickness Due to Infrared Radiation

The material should exhibit a limited “thickness effect”, meaning that the contribution of radiation to heat transfer should be minimal, resulting in only a slight variation in thermal conductivity with thickness. Generally, infrared radiation is absorbed more as a material’s density increases. This phenomenon is particularly evident in thin, low-density specimens. For example, in EPS with a thermal conductivity of 0.046 W/(m·K), the effect of infrared radiation is 10% at a thickness of 20 mm and 5% at 50 mm.

The standard EN 12939 [11] and the document CEN/TR 131 [12] describe the effect of thickness on thermal conductivity measurement. Figure 1 illustrates the relationship between thermal resistance and specimen thickness for various thermal insulation materials. The extrapolation to zero thickness, R₀, of the straight portion of the graph depends on material properties and testing conditions [12]. The thickness $d_{\mathit{\infty }}$ indicates the beginning of a straight portion of the plot of thermal resistance.

The thermal resistance, R, of a flat specimen of low-density material may be expressed as follows:

$$R=R_0^{\prime }+d/{\lambda }_t,$$

(1)

where $R_0^{\prime }$ is not necessarily independent of the thickness d, and the following is true:

$${\lambda }_t={\lambda }_{cd}+{\lambda }_r,$$

(2)

where

• λ_t is the thermal transmissivity;
• λ_cd is the combined gaseous and solid thermal conductivity;
• λ_r is the radiativity.

Above the thickness limit, $d_{\infty }$, the ratio Δd/ΔR is constant; the thermal transmissivity λt, which is an intrinsic material property independent of experimental conditions, can be measured. In this case, the radiativity, λ_r, and the gaseous and solid thermal conductivity, λ_cd, can also be defined as material properties and expressed as λ_t = λ_cd + λ_r. Nevertheless, the transfer factor T = d/R is not yet independent of the thickness d. For a thermal insulation material, the criteria this region is to have T and λ_t which do not differ from one another by more than 2% [10,13,14].

The thickness effect can be determined by the following value:

$$(1-L)$$

(3)

where

• L is the ratio between the transfer factor and thermal transmissivity given by L = T/λ_t.

For mineral wool and EPS, according to the standard EN 12939 [11], the thickness effect may be considered not relevant if (1 − L) ≤ 0.01. If the (1 − L) is above 0.01, an assessment of the thickness effect should be performed.

According to EN 12939 Table 1, for mineral wool with a thermal conductivity below 0.035 W/(m·K), the thickness effect is not relevant ((1 − L) < 0.01) for products having a thickness above 32 mm. The effect is lower than 2% ((1 − L) < 0.02) when the thickness is above 20 mm.

According to EN 12939 Table 3, for expanded polystyrene with a thermal conductivity below 0.032 W/(m·K), the thickness effect is not relevant for a product having a thickness above 70 mm. The effect is less than 2% when the thickness is above 34 mm.

A study made by Blazejczyk [15] shows that, for white EPS having a thermal conductivity of 0.040 W/(m·K) and a density of 16.9 kg/m³, the thickness effect is lower than 2% for thicknesses above 40 mm. On grey EPS having a thermal conductivity of 0.0314 W/(m·K) and a density of 14.6 kg/m³, the thickness effect is lower than 2% for thicknesses above 7 mm.

A study made at LNE on three materials, two white EPS, one with a density of 21.2 kg/m³ and one with a density of 31.8 kg/m³, and one stone wool with a density of 41.1 kg/m³, was conducted to determine the effect of thickness on thermal conductivity. The procedure was to measure several stacks of panels and measure their thermal conductivities in terms of function of thickness. The influence of thickness on the low-density EPS was observed below 57 mm and, on high-density EPS, the influence was not seen above 30 mm (Figure 2 and Figure 3). In stone wool, the influence is not seen above 27 mm (Figure 4).

The

Table 4. Minimum and maximum allowed specimen thickness.

	ρ (kg/m3)	${\mathit{d}}_{\mathit{\infty }}$ (mm)
EPS	21.2 31.8	57 below 30
MW	44.1	below 23

summarizes the influence of density on the thickness $d_{\mathit{\infty }}$ above which the ration Δd/ΔR is constant.

2.3. Homogeneity and Density

The EN ISO 17034:2018 standard requires the reference material producer to determine the homogeneity of every property of the batch, which should be quantified as an uncertainty contribution to the certified value.

For mineral wool and expanded polystyrene, the thermal conductivity as a function of density can be modeled by the following relationship:

$$\lambda \left(\rho \right)=A+B\rho +{\displaystyle \frac{C}{\rho }}$$

(4)

where the coefficients A, B, and C represent, respectively, the gaseous, the solid, and the radiative parts. An example of the curve λ(ρ) is given in Figure 5 which was determined with the coefficients listed in

Table 5. Example of parameters A, B, and C of the λ(ρ) curves for mineral wool and expanded polystyrene [16,17].

	MW	EPS
A (W/(m·K))	0.02490107	0.025314
B (W·m2/(kg·K)	6.23 × 10−5	5.17 × 10−5
C (W·kg)/(K·m4)	0.2284	0.1736

.

For densities below 50 kg/m³, the variation in thermal conductivity is significant and can lead to a larger uncertainty due to the density variation among specimens. Above 50 kg/m³, the evolution can be given by the coefficient B, which corresponds to the slope of the curve, and can be easily used in the uncertainty budget. This means a thermal conductivity of above 0.030 W/(m·K). Unfortunately, only mineral wool product are available on the market for these densities with a suitable thermal conductivity level. Additionally, the deformation of the material under a specified load should be taken into account. A higher density reduces the deformation.

2.4. Other Characteristics

The length and width of samples should be in accordance with the dimensions of the available apparatuses in Europe, which are 200 × 200, 300 × 300, 500 × 500, 600 × 600, 800 × 800, and 1000 × 1000. Most of the devices installed in Europe are dedicated to measuring specimens of 300 × 300 mm and 600 × 600 mm.

To ensure as low a contact thermal resistance as possible between the plates of the lambda-meter, the roughness of the surfaces should be low. The surface of the specimen should be smooth or sanded to remove the asperities. The hardness of the specimen should be hard, in case it needs to be sanded.

Last but not least, the material should be manufactured by a company within a reasonable time. Special orders may induce unacceptable costs.

2.5. Investigation of Standard and Alternative Materials

Table 1 lists the major existing thermal insulation materials. This paragraph makes a comparative study on the thermal conductivity value linked with long-term stability and the effect of the environment. The key characteristics of a reference material are listed in

Table 6. Summary of reference material key characteristics.

Characteristics	Criteria	Procedures
long-Term stability of thermal performance	resistance to environmental factors such as temperature fluctuations, humidity, and aging.	using materials known for their durability. accelerated aging tests and real-time aging studies to simulate long-term use
thermal conductivity range	thermal conductivity values that match or closely resemble existing European products.	selection of material laboratory measurements using standardized methods (e.g., ISO 8301 [13], ISO 8302 [14]).
Minimal Dependence on Density	low correlation between density and thermal conductivity.	testing samples with varying densities and analyzing the relationship between density and thermal conductivity
appropriate thickness	thicknesses within a specified range to ensure compatibility with measurement devices.	producing and testing samples at different thicknesses to determine optimal ranges

. This table provides the criteria for each characteristic and the procedures to obtain them.

2.5.1. Long-Term Stability

The aging methodology of thermal insulation material is defined in the harmonized standard for cellular plastic, EPS, XPS, PU, and PF, which have a blowing agent whose thermal conductivity is below that of air, like pentane HFC, CO2, or, now, HFO, which stay in the products for a long time period. These products present a thermal conductivity of below 0.030 W/(m·K), but they increase with time due to diffusion of air into the product. The aging methods give a prediction of the time-averaged aged value over 25 years for thermal conductivity. For PU, the product is placed in an oven at 70 °C for 6 months. For PF, the temperature is set to 110 °C for 2 weeks or 70 °C for 6 months, followed by conditioning at 23 °C/50% relative humidity. For XPS, the product is sliced into several layers 10 mm thick, and conditioned for 30, 50, or 90 days, depending on the product thickness, at 23 °C and 50% RH of moisture. In these types of products, the increase in thermal conductivity can exceed 0.007 W/(m·K).

In organic products, aging is influenced by moisture and temperature, but the harmonized standards do not provide any specific assessment methods. Therefore, laboratories use different aging procedures based on temperature and moisture cycling (dry-heat, humid-heat). The impact on thermal performance is not sufficiently documented to draw conclusions on the aging of organic products.

2.5.2. Effect of Environment

A study [18] was conducted on wood fiber products with densities ranging from 60 to 130 kg/m³ conditioned in different environments. The increase in moisture content from a dried state (specimen dried at 70 °C until reaching mass constant) to a wet state (specimen conditioned at 23 °C with 50% RH and 95%RH) resulted in a water intake of from 7 to 20% of the total mass (Figure 6). This moisture increase led to a rise in thermal conductivity of from 2% to 10% compared to the dried state (Figure 7).

2.6. Summary of the Evaluation of Standard and Alternative Materials

Table 7. Investigation of several existing material to become reference material.

Product	Long Term Stability	Moisture Influence on λ (Lab Condition)	Range of Lambda W/(m·K) (after Ageing Procedure)	Dependence on Density	Dependence on Thickness	Dependence on Applied Load
MW	No ageing when stored at laboratory condition	_	λ ≥ 0.030	yes, λ(ρ)	effect of IR on for low density product	yes (load change density)
EPS		_	λ ≥ 0.030
CG		_	λ ≥ 0.035	yes		no
PU	Ageing	_	λ ≥ 0.021		λ depend of the thickness product	no
XPS		_	λ ≥ 0.027 or λ ≥ 0.030 depending on the blowing agent			no
PF		Effect of moisture	λ ≥ 0.021			no
PEB	require further studies	Effect of moisture	λ ≥ 0.050			no
WF			λ ≥ 0.035	yes	require further studies	yes (load change density)
Sheep wool
CI			λ ≥ 0.038
Cotton			λ ≥ 0.038
Hemp			λ ≥ 0.06
Straw			λ ≥ 0.05
ICB			λ ≥ 0.040			no
WW			λ ≥ 0.070
Aerogels binder	Sensible to moisture	Effect of moisture	λ ≥ 0.015		yes	yes
VIP	Ageing Air and moisture transfer	_	λ ≥ 0.008		yes	No

summarizes the behavior of various products in terms of long-term stability, environmental influence, and the dependence of thermal conductivity on density thickness and applied load.

All bio-based materials, such as wood fiber (see Section 2.5.2), are sensitive to moisture, with their thermal conductivity increasing as moisture levels rise. All cellular plastics undergo aging over time (see Section 2.5.1), resulting in a gradual increase in thermal conductivity. The aging of vacuum insulation products is primarily caused by air penetrating the product through its envelope, similarly leading to an increase in thermal conductivity over time, much like in cellular plastics.

Based on these findings, two materials, EPS and MW, could be potential candidates for reference materials. Aerogel may also become a viable solution in the future if its long-term stability and low sensitivity to moisture can be demonstrated.

3. Manufacturing Specification

The JRC Geel has produced two reference materials for thermal conductivity since 2000: the IRMM 440 in 2000 and the ERM FC 440 in 2021. These two reference materials are made of glass wool fibers. The main specifications for these two materials are as follows:

• Low thermal conductivity, below 0.035 W/(m·K), to be in line with currently used insulation materials;
• Sandable, to be able to remove the tracks from the conveyor belt from processing and to generate a plane surface;
• A density of around 80 kg/m³. This value is a compromise between the minimum of the density–conductivity curve, the hardness of the material, and the capacity to be sandable. Thus, the variation in density would have a predictable effect on the thermal conductivity;
• Thickness of about 35 mm.

According to the specifications above and the production possibilities, both reference materials were defined in

Table 8. Characteristics of both European reference material.

	IRMM 440	ERM FC 440
Size of boards: at least	1.200 m × 1.200 m	2.400 m × 1.200 m
Thermal conductivity (W/(m·K))	Below 0.0035	Below 0.0035
Nominal thickness: mm	35	30
Flatness tolerances: 0.4 mm	0.4 mm
Nominal density (kg/m3)	70	135
Compressive stress at 10% deformation (kPa)	_	30
Range around nominal density between boards	±5 kg/m3	±2.6%
Maximum thickness effect of thermal conductivity value	0.5%	0.26%
Surfacing of the specimen (sandpapered)	yes	yes
Size of the specimen (mm)	IRMM 440A 300 × 300 IRMM 440B 500 × 500 IRMM 440C 600 × 600 IRMM 440D 600 × 600	ERM FC 440a 300 × 300 ERM FC 440b 500 × 500 ERM FC 440c 600 × 600

.

In comparison, NIST, in the USA metrological laboratory (National Institute of Standards and Technology), has produced several reference materials [5] that are mainly made of fibrous glass board. Since 1978, NIST has developed three thermal insulation reference material SRM’s: fibrous glass board, 1450 1450a, 1450b, 1450c, and 1450d [19,20]; fibrous glass blanket, 1451 and 1452; and fumed-silica board, 1449.

All existing reference materials were made for plate equipment, GHP or HFM. At this time, there is no certified reference material for pipe measurement according to ISO 8497 [6,21].

4. Qualify Reference Material

The reference material should be qualified by means of a guarded hot-plate round-robin test or intercomparisons between European laboratories. JRC Geel requires at least five metrological or notified laboratories to qualify a reference material. The certified thermal conductivity is then determined by the mean of the measured values by each laboratory, and the uncertainty is given by the cumulated uncertainty of each apparatus.

The thermal conductivity measurements are carried out by a Guarded Hot Plate (GHP), which is an absolute device. The guarded-hot-plate method, which has been standardized by ISO 8302 [14] or EN 12667:2001 [10] standards, determines the steady-state thermal transmission properties of flat slab specimens with a low thermal conductivity. The standard test methods for the guarded hot plate utilizes the one-dimensional steady-state thermal conductivity equation for the determination of thermal conductivity (λ):

$$\lambda ={\displaystyle \frac{QL}{2A\mathsf{\Delta }T}}$$

(5)

where Q is the time-rate of one-dimensional heat flow through the meter area of the guarded hot plate in Watts; A is the metering area of the apparatus normal to heat flow in square meters; ΔT, in Kelvin, is the temperature difference across the specimen hot (T_h) and cold surfaces (T_c); and, L, in meters, is the in situ thickness of the pair of specimens. Values of λ are reported at the mean temperature T_mean given as follows:

$$T_{mean}=\left(T_h-T_c\right)/2$$

(6)

Each laboratory gives their expanded uncertainty (U) with a level of confidence of 95% with a coverage factor of k = 2 [22]. The relative expanded uncertainty, U_r (%), in Table 8 is defined as U/|λ| and was determined by each laboratory.

The participant laboratories for the qualification of IRMM 440 and ERM FC400 are listed in

Table 9. Laboratories participating in the qualification of both European Reference material.

IRMM 440	ERM FC 440
LNE (France)	LNE (France)
DFT Italy(DFT)	MPA-NRW (Germany)
EMPA (Swiss)	MPA Stuttgart (Germany)
FIW (Germany)	FIW (Germany)
NPL (UK)	IMBiGS (Poland)
Rise (Sweden)

. All the chosen laboratories are metrological laboratories in Europe and are accredited by an Iliac member. The devices used for the measurements are listed in

Table 10. Characteristics of the different GHPs used for the determination of IRMM 440.

	Laboratory
Parameter	1	2	3	4	5	6
Plate, mm	610 × 610	300 × 300	750 × 750	800 × 800	610 × 610	400 × 400
Meter plate, mm	300 × 300	148 × 148	300 × 300	300 × 300	202 × 202	202 × 202
Guard width, mm	155	76	350	250	150	99
Guard Gap	1	2	2	1.5	2	3
Plate emittance	0.86 ± 0.05	0.89	0.92	0.93	0.89	0.904
Operation mode	Two specimens	Two specimens	Two specimens	Two specimens	Single specimen	Single specimen
Ur (%), (k = 2)	±1.0%	±0.7	±1.0	±0.95	±1.4	±0.7

and

Table 11. Characteristics of the different GHPs used for the determination of ERM FC440.

	Laboratory
Parameter	1	2	3	4	5
Plate, mm	610 × 610	800 × 800	800 × 800	304 × 305	500 × 500
Meter plate, mm	300 × 300	400 × 400	400 × 400	150 × 150	300 × 300
Guard width, mm	155	200	200
Guard Gap	1	2	2	2	1.5
Plate emittance	0.86 ± 0.05	0.97	0.95	0.903	>0.9
Operation mode	Two specimens	Two Specimens	Two Specimens	Two specimens	Two specimens
Ur (%), (k = 2)	±1.0%	±1.0%	±1.0%	±1.0%	±1.0%

. All guarded hot plates were calibrated in accordance with EN 1946-2 [23].

The thermal conductivities were measured between −10 °C and 50 °C or 70 °C, depending on which reference material was used, IRMM 440 or ERM FC 440. The evolution of thermal conductivity with the temperature is close to a straight line for both materials, which is what is expected in this range of temperatures. In Europe, the thermal conductivity of insulation materials for building insulation is mainly determined at 10 °C. The ERM FC440 shows a thermal conductivity higher than the IRMM 440 due to its higher density. However, the ERM FC400 is also thinner than the IRMM 440, which may fall below the specification of some lambda meters.

Table 12. Characteristics of both reference materials.

	IRMM 440	ERM FC 440
Density (kg/m3)	64–78	137–133
Thickness (mm)	35	28
Range of temperature (°C)	−10 ≤ θ ≤ 50	−10 ≤ θ ≤ 70
Certified value (W/(m·K))	$2.93949\times {10}^{-2}+1.060\times {10}^{-4}\theta +2.047\times {10}^{-7}$	$0.03104+1.1\times {10}^{-4}\theta$
Uncertainty (W/(m·K)) or %	0.00028	1.1%

lists the main characteristics and Figure 8 shows the thermal conductivity of both materials as a function of temperature.

The evolution of thermal conductivity over time was also evaluated by two laboratories over a period of 14 years (Figure 9) [24]; no evolution of thermal conductivity was seen. Between August 2000 and October 2013, another laboratory conducted more than 100 tests at 10 °C on various Heat Flow Meter apparatuses. The deviations from the laboratory mean ranged from 1% to −1.5% (Figure 10) [24]. Additionally, a regression analysis of these data over time showed no significant trend at the 95% confidence level.

5. Consequences on the European Market and Next Generation of Reference Material

The organization of the European market is very specific compared to other countries around the world. In Europe, since 2022, CE marking has been mandatory to place a product on the market, and the thermal resistance of thermal insulation products must be determined by a third-party organization called a “notified body”. The notified body is accredited by a member of the European Co-operation for Accreditation (EA) and is approved by the authorities of each member state. Moreover, the standard EN 13172 makes it mandatory to use the European reference standards (IRMM 440 or ERM FC 440). All thermal values are then traceable to SI (International System of Units) based on the same reference material for the apparatus’s lambda-meter in Europe.

When we compare the results of round-robin tests organized in two different markets, Europe and North America, the differences are significant. To perform a fair analysis, we need to take into account that the best uncertainty a guarded hot plate can achieve is 1% and 3% for a Heat Flow meter. Within Europe, several round-robin tests have been organized by the Keymark scheme for the voluntary certification of insulation since 2001. The analyses were carried out on three materials: EPS, MW, and IRMM 440. Twenty-five laboratories representing 65 pieces of equipment participated in these measurements. The maximum deviation observed since 2021 was 2.7%. Moreover, for EPS, 98.7% of measurements, and, for MW, 96% of measurements, are within ±1.5% of the mean values [25,26,27].

In North America, 30 round-robin tests were organized between 1986 and 2004, involving 37 laboratories. The compilation of all these measurements shows a maximum deviation between laboratories of 12% of the grand mean [28]. Moreover, on the last 19 rounds, 72% of measurements were within ±2% and 95.6% were within ±6%. Nevertheless, these discrepancies are not related to the materials themselves but to the overall organization of the system. Indeed, when intercomparisons are made between NIST and LNE, two metrological laboratories, one North American and the other European, the difference between the two laboratories does not allow for statistical discrimination with a difference lower than 1% between them [29].

This comparison highlights the significant impact of the European organization, including the reference materials and the CE marking, on thermal conductivity measurements. The system enforced in Europe has improved the competencies of the accredited laboratories and the quality of manufacturers’ laboratories. Moreover, the lower uncertainty seen in Europe has pushed the manufacturers to enhance their products, as the competition between them is within a percent of thermal conductivity: approximately 0.0003 W/(m·K) (Figure 11).

The thermal conductivity value is also crucial for calculating the energy performance of buildings. If the thermal resistance of the installed product is known with an uncertainty of 1.5% or 12%, this uncertainty will be propagated to the energy performance of the building.

To conclude this section, we observe the evolution of the insulation materials market in two directions. The first direction involves increasing the installed thermal resistance by either increasing the thickness of the installed products or reducing the thermal conductivity of the product, for example, by using superinsulation materials (SIMs) or improved cellular plastics. SIMs have a thermal conductivity ranging from 0.005 W/(m·K) for vacuum insulation materials to 0.015 for products with aerogel, while some plastics can achieve a thermal conductivity of 0.002 W/(m·K). The second direction is reducing CO₂ emissions by utilizing bio-based materials or materials that incorporate bio-based components. These materials exhibit a thermal conductivity of around 0.004 W/(m·K). This market evolution necessitates the development of two additional types of reference materials: one with lower conductivity and one with a thickness of around 100 mm.

6. Conclusions

Reference materials (RMs) and certified reference materials (CRMs) play a pivotal role in ensuring the accuracy and consistency of thermal conductivity measurements for insulation materials in Europe. Adhering to ISO 17034:2018 standards, these materials provide a benchmark for calibration, quality control, and performance checks in manufacturing and testing environments. The rigorous specifications for long-term stability, appropriate thermal conductivity range, and minimal variation due to density and thickness underline the importance of selecting suitable reference materials.

The process of defining a reference material is lengthy and involves several key factors. These include ensuring the long-term stability of thermal performance, aligning the level of thermal conductivity with products produced in Europe, minimizing the dependence of thermal conductivity on density to reduce uncertainty, and ensuring the thickness is suitable for lambda meters. The production of reference materials requires high-quality manufacturing and testing by at least five European metrological laboratories. The certified values are then determined by averaging all laboratory results and accounting for combined uncertainties. Once established, reference materials become essential for all stakeholders, including test laboratories, manufacturers, and accredited bodies. The European market’s organization hinges on three pillars: reference materials, CE marking, and notified bodies.

The European market, regulated by CE marking and EN standards, benefits from the use of standardized reference materials like IRMM 440 and ERM FC 440. These materials enable reliable comparisons across different laboratories and testing equipment, thereby enhancing the overall quality and competitiveness of insulation products. This framework has significantly improved measurement quality, with 96% of laboratories achieving results within ±1.5% of the mean value. The comparison with the North American market underscores the effectiveness of the European approach in reducing measurement uncertainties and improving product standards.

Looking ahead, the insulation industry faces the dual challenge of enhancing thermal resistance and reducing CO₂ emissions. The importance of thermal insulation in buildings for energy consumption and comfort is well-recognized, leading to regulations since the late 20th century to enforce building insulation. The EU Green Deal, introduced in 2020, further aims to enhance the energy performance of buildings. This necessitates the development of new reference materials with lower thermal conductivity and greater thickness, catering to the evolving market demands for superinsulation materials and bio-based alternatives. The continued evolution of reference materials is critical for supporting the industry’s goals of improving energy efficiency and sustainability in building insulation. Developing these new reference materials presents the next challenge for metrological laboratories, ensuring continued progress in the insulation industry and contributing to improved energy efficiency and sustainability in building practices.