APPLICATION NOTE

How to Measure Thin Insulation Samples

with the HFM
Fabia Beckstein

1 HFM Lambda Small and HFM Lambda Medium

NGB · Application Note 217 · EN · 0721 · Technical specifications are subject to change.

Introduction

Known insulation materials like mineral wool or polymer U-Value

foams are usually produced with high thickness (several
centimeters) to meet the required U-value for the ther- The U-value reflects the heat flow through a com-
mal insulation of buildings. A suitable measurement ponent dependent on the temperature gradient
device for determination of the thermal conductivity between the wam and cold side of a unit [W/(m2·K)].
(λ) is the HFM 446 Lambda Medium (figure 1). How- The unit describes the energy flowing through
ever, insulation materials are also applied in other areas 1 square meter due to a temperature difference of 1 K.
with other thicknesses, for example, in the thermal and This value characterizes the insulation properties of
sound insulation of floors. The thickness of such insula- a component; this means in practice, the lower the
tion materials is often only a few millimeters. The follow- U-value, the better the insulating effect. The higher
ing measurements show how such thin materials can the U-value, the poorer the insulating effect. The
successfully be investigated with the HFM 446 Lambda building then loses more heat on cold winter days.
Medium.

1 |5 NETZSCH-Gerätebau GmbH
Wittelsbacherstraße 42 ∙ 95100 Selb ∙ Germany
Phone: +49 9287/881-0 ∙ Fax: +49 9287/881505
at@netzsch.com ∙ www.netzsch.com
 APPLICATIONNOTE How to Measure Thin Insulation Samples with the HFM

Measurement Method

A temperature gradient is defined between two plates λ Thermal conductivity [W/(m∙K)]

by a material to be measured. By means of two high- d Thickness [mm]
precision heat-flow sensors in the plates, the heat flow R = d/ λ Thermal resistance [m2∙K/W]
into the material and out of the material, respectively, U = 1/R Heat transfer coefficient [W/(m2∙K)]
is measured. Once equilibrium of the system is reached
and the heat flow is constant, the thermal conductivity
can be calculated with the help of the Fourier equation
and knowledge of the measurement area and thickness
of the sample (see schematic figure 2).

Measurement Conditions
∙ Stacking of samples, mentioned in DIN EN 12667

∙
A natural fiber insulation board with a thickness of 4 mm
was investigated. The thermal resistance (R = d/λ) of Measurement of one sample with additional external
such thin samples presents a challenge for measure- thermocouple and interface layers (=instrumentation
ment. Samples with thermal resistance lower than kit), described in DIN EN 12664 for samples with a
approximately 0.5 m²∙K/W cannot be measured with thermal resistance of < 0.5 m2∙K/W.
the HFM as a standard measurement (DIN EN 12667).
The contact resistance between the plates and sample The measurements were carried out at a mean sample
is no longer negligible and will affect the result. To over- temperature of 25°C. The temperature difference
come the issue of low thermal resistance, measure- between the plates was 20 K. The pressure on the sample
ments with two different approaches were performed: was approximately 2 kPa.
NGB · Application Note 217 · EN · 0721 · Technical specifications are subject to change.

Thickness gauge
Load

Load

Hoisting device

Upper heat sink

Peltier system
Electronics and data
acquisition system

Hot plate
Heat flux transducer

Test sample Direction of

heat flow
Heat flux transducer

Cold plate
Peltier system
Lower heat sink

Cooling system

2 Schematic of the HFM which is delivered calibrated

2 |5 NETZSCH-Gerätebau GmbH
Wittelsbacherstraße 42 ∙ 95100 Selb ∙ Germany
Phone: +49 9287/881-0 ∙ Fax: +49 9287/881505
at@netzsch.com ∙ www.netzsch.com
 APPLICATIONNOTE How to Measure Thin Insulation Samples with the HFM

Stacking of Samples With a thickness higher than 20 to 24 mm (5 to 6 lay-

ers), the thermal conductivity is constant and no longer
Figure 3 shows the thermal conductivity versus the overall dependent on thickness. This is the region where the
thickness of stacked samples (1 to 8 layers). The measure- contact resistance is negligible and measurements can
ment data is summarized in table 1. be considered reliable. The thermal resistance of the
sample is higher than approximately 0.5 (m²∙K)/W.
In the low thickness range, the thermal conductivity
shows a dependence on thickness. The contact resis-
tance between the sample and HFM plates affects the
result (reduced thermal conductivity).

3 Thermal conductivity of a natural fiber insulation board versus thickness (= number of layers)
NGB · Application Note 217 · EN · 0721 · Technical specifications are subject to change.

Tab 1. Measaurement results of the stacked samples of a 4-mm thick naterual insulation fiber board

Thickness Thermal Conductivity* Thermal Resistance

Number of Layers
[mm] [W/(m∙K)] [(m2∙K)/W]

1 4 0.04214 0.0958

2 8 0.04447 0.1812

3 12 0.04565 0.2582

4 16 0.04697 0.3387

5 20 0.04745 0.4214

6 24 0.04779 0.5021

7 28 0.04749 0.5906

8 32 0.04734 0.6757

* all results ± 3%

3 |5 NETZSCH-Gerätebau GmbH
Wittelsbacherstraße 42 ∙ 95100 Selb ∙ Germany
Phone: +49 9287/881-0 ∙ Fax: +49 9287/881505
at@netzsch.com ∙ www.netzsch.com
 APPLICATIONNOTE How to Measure Thin Insulation Samples with the HFM

Figure 4 (thermal resistance over thickness) confirms that Therefore, the thickness (= distance between the exter-
the measurements with stacked samples are reliable. nal thermocouples) is not precisely known. Due to the
The thermal resistance increases linearly with increasing low thickness of only 4 mm, even a low level of pene-
thickness. The linear trend line gives an R² fit of 0.99972 tration can cause a high deviation in the result (rela-
and the slope is an indicator for the thermal conductiv- tive error in thickness causes the same relative error in
ity (slope m = R/d = 1/λ → λ = 0.04855 W/(m∙K)). This thermal conductivity).
value is in good agreement with the measurement
results of the stacked sample with a thickness higher Table 2 shows the results of a measurement with the
than ~20 mm; see table 1. instrumentation kit. The measurement of one layer with
the instrumentation kit produces a value that is approx-
Instrumentation Kit imately 10% higher than the results from the stacked
samples. This 10% increase in the thermal conductivity
For low thermal resistance samples, measurements with value is most probably caused by a thickness value that
the instrumentation kit (= external thermocouples and is incorrect by 10% due to penetration by the external
interface layers) can also be a good solution. The issue of thermocouple (200 μm on each side). This is confirmed
contact resistance is solved by direct measurement of the by the measurement of 1 and 2 layers with the instru-
surface temperature. For stiff samples, the instrumenta- mentation kit and calculation of the thermal conductivity
tion kit is a good choice. Since the 4-mm thick natural with adjusted thickness (= thickness minus 2 x 200 μm).
fiber insulation board ist not completely stiff but still flex- The thermal conductivity with adjusted thickness is in
ible, another source of uncertainty exists. The external good agreement with the values from the measure-
thermocouples can penetrate into the sample’s surfaces. ments with stacked samples.
NGB · Application Note 217 · EN · 0721 · Technical specifications are subject to change.

4 Thermal resistance of a natural fiber board versus thickness (= number of layers)

Tab 2. Measurement results of the 4-mm thick natural fiber insulation board with instrumentation kit

Thermal Conductivity with

Thickness Thermal Conductivity* Adjusted Thickness
Number of Layers Adjusted Thickness
[mm] [W/(m·K)] [mm]
[W/(m∙K)]
1 4 0.05281 3.6 0.04753

2 8 0.05071 7.6 0.04817

4 |5 NETZSCH-Gerätebau GmbH
Wittelsbacherstraße 42 ∙ 95100 Selb ∙ Germany
Phone: +49 9287/881-0 ∙ Fax: +49 9287/881505
at@netzsch.com ∙ www.netzsch.com
 APPLICATIONNOTE How to Measure Thin Insulation Samples with the HFM

Summary sufficient thickness. Measurement with external thermo-

couples (instrumentation kit) on flexible samples can
The thermal conductivity of thin and flexible insulation result in falsely inflated thermal conductivity values due
materials can be measured with the HFM 446 Lambda to the possible penetration of the thermocouples into
Medium by stacking several layers of the material to a the sample’s surface.
NGB · Application Note 217 · EN · 0721 · Technical specifications are subject to change.

5 |5 NETZSCH-Gerätebau GmbH
Wittelsbacherstraße 42 ∙ 95100 Selb ∙ Germany
Phone: +49 9287/881-0 ∙ Fax: +49 9287/881505
at@netzsch.com ∙ www.netzsch.com