Method and Apparatus for Comparing Infrared Reflection Property of Insulating Panels

This paper discloses the method and apparatus we are using to measure and compare the amount of infrared energy reflected directly from the surface of various insulation panels.

Section A-A

In response to customer interest, we have constructed a test station that can measure and compare the amount of reflected infrared energy to a standard reflector. The station consists of a tubular steel stand on which is mounted an insulated chamber constructed of 1” thick GemcoWool® board. The side walls are slotted to hold 6 heat lamps. There is a board on top that has a fixed opening. Samples being measured lie on this board and completely cover the opening.

Two ½” thick board (Slit Aperture Board) made from ½” GemcoLite® are positioned to prevent direct exposure of the thermocouple from the lamps.

The thermocouple is mounted to an adjustable bracket which is in turn mounted to a thermocouple support made of tubular steel, and affixed to the bottom tray of the stand. RTD thermocouples are used for their accuracy.

Power to the lamps is supplied by a Watlow solid state power supply. Power is controlled by an Altec loop calibrator. Power consumed is measured by an AC watt transducer. Data is recorded on a data logger.



The design of the stand limits the energy gain at the exposed thermocouple to the radiant energy reflected from the sample and support structure, along with any emitted energy from the surfaces and conduction from the warming air. It is assumed that the energy delivered to the thermocouple from everything except the sample, is independent of the sample being tested, and repeatable.

  • Baseline 0
    • We record energy gain at the exposed thermocouple with no sample in place. The energy density gain measured in this manner is assumed to be inherent in the test stand. We establish this as a baseline for 0% reflection.
  • Baseline 100
    • We record energy gain at the exposed thermocouple with a “standard” sample in position. In our use of the station so far, we have used Quartzel® with an advertised reflection of 90%. We subtract the baseline 0 value from the energy gain, and then normalize the values to 100% by dividing by 0.9. See the notes about changing from Quartzel® to Gold for establishing the Planned Baseline 100 Changes section later.
  • Comparisons
    • We record energy gain at the exposed thermocouple with various other insulating panels. We calculate a comparative percent reflection by subtracting the baseline 0 value from the measured energy gain, then finding the percent energy gain relative to our established percent scale.



The sample to be measured is cut to approximately 9”x12” so it will cover the opening in the sample holding board. The fixed opening guarantees we will have the same area of each sample exposed to the heat lamps.

The sample is examined for flatness, and inspected and cleaned if necessary.

All doors to the lab remain closed during the test to help prevent stray drafts. In early testing a shroud was placed around the test stand, but this has proven unnecessary. Ceiling fans and any other source of draft are turned off for the test session.

The data logger is started, then after a second or two, the power is turned on to the lamps. Note that the loop calibrator has been preset, and remains at the same level for the test. The power is left on for just over 2 minutes, then the power is turned off and the data logger stopped.

The data is moved to a pc and analyzed. It is easy to see in the data the exact moment the power is turned on. The data used to calculate the energy density gain at the thermocouple begins at this moment and continues for 65 seconds from this point. The energy gathering surface is limited to the area of the thermocouple, and is small at just under 0.03125 in^2, so the rate of energy transfer is not large. Data is sampled at 100 samples per second. Averages of each 5 second period are calculated for a total of 13 time segments, 0 to 4.99 seconds, 5 to 9.99 seconds, etc. through 60 to 64.99 seconds.

Data averaged includes

  • Ambient temperature
  • Exposed temperature
  • Power consumed

For each 5 second period, the temperatures are converted to Kelvin and energy flux (density) is calculated using the Stefan-Boltzman constant. The energy density gained by the thermocouple is calculated in W/m^2. This gain is expressed as a percentage of the amount of power supplied to the lamps by dividing the energy density gain (in joules) by the energy supplied to the lamps (also in joules). The gain is adjusted by subtracting the gain expected from the baseline 0 data.

This percentage of energy transferred is what is compared with the baseline values, and used to calculate the final comparative percentage.

A final value used for quick comparisons is found by averaging the reflection percents calculated for all 5 second intervals.

Sample Sessions

The comparison data for typical test sessions for three materials are shown in the following tables.

Using GemcoLite® as the reflective surface sample is shown in this table.

Gemcolite® Table

Using Quartzel® as the reflective surface sample is shown in this table.

Quartzel® Table

Using Q-Board™ as the reflective surface sample is shown in this table.

Q-Board™ Table

Using AmorSil™ as the reflective surface sample is shown in this table.

Armorsil Table

Planned Baseline 100 Changes

We initially decided to use Quartzel® fused quartz fiber as the material to establish the 100% baseline values. This material, manufactured by Saint-Gobain, is advertised to have high reflective properties in IR, and generally considered to have the best infrared reflection capabilities of the insulating panel materials available. The samples we are using are flame polished rigid panels. The Baseline 100 values we established from this material have caused me some concern. Our new Q-Board, which we are specifically developing to have IR reflection properties competitive with Quartzel®, was intended to be “as good as”. However, we have seen comparative values of Q-Board that exceed 100%, as shown in the Sample Sessions data example above. This indicates the Quartzel® samples we have used to establish the 100% baseline values are not reflecting as much IR as we expected. I have concluded that variations in the manufacturing process are the cause. The particular Quartzel® samples I am using do not seem to be reflecting a full 90% of the IR. (Quartzel® is a registered trademark of Saint-Gobain Quartz, a business unit of the Saint-Gobain group).

We have decided to change to a known material to establish the 100% values. I considered using Aluminum foil, however, Aluminum’s IR reflection properties vary considerably with wavelength. Gold is far more consistent, and has well document very high reflective properties in IR.

We are in the process of acquiring a gold mirror of suitable size to use for calculating more accurate 100% baseline values.


Work with the test station to date has shown good repeatability. We have considered increasing the exposed thermocouple area so it would catch more energy, but this may not be required.

The information we are acquiring is going to be of great value in helping our customers make informed decisions when comparing insulating panels for their infrared reflection properties.

Greg Gorby
Research and Development Manager
Refractory Specialties, Inc.
Specialty Ceramics Inc.