ASTM D2149: Dielectric Constant

Reviewed by SEMI Gr

ASTM D2149 measures these properties at frequencies up to 10 MHz and temperatures up to 500 °C. This temperature range is relevant to many technical ceramic applications, where ceramics are used because polymeric insulating materials may not withstand the operating conditions.

Room-temperature dielectric data only tells you how a component behaves under laboratory conditions. In actual use, an insulator in a downhole tool, furnace feedthrough, engine sensor, or high-power RF assembly may operate at much higher temperatures. As the temperature changes, both permittivity and dielectric loss can change as well.

MatX Lab performs high-temperature dielectric measurements on technical ceramics, glass-ceramics, and inorganic insulating materials.

Scope of the Method

The method covers solid dielectric materials measured across two variables: frequency (up to 10 MHz) and temperature (up to 500 °C).

Ceramics are the usual subject. Alumina, steatite, cordierite, aluminum nitride, glass-ceramics, and porcelain all serve as insulators where operating temperature rules out organic materials, and all change electrical behavior as they heat.

The quantities are the same as in ambient dielectric work: relative permittivity and dissipation factor, measured through the capacitance and loss of a specimen held between electrodes.

What Changes as a Ceramic Heats

Three mechanisms drive the temperature dependence, and they act in different directions.
Ionic mobility rises with temperature. Alkali ions in a glassy grain-boundary phase become mobile, contributing to conduction and raising the dissipation factor, often steeply above a few hundred degrees.
Dipolar and interfacial polarisation shift with temperature, which moves relaxation peaks through the measured frequency range. A loss peak sitting outside your frequency band at 25 °C can move into it at 300 °C.

For many ceramics, permittivity increases gradually as the temperature rises. Ferroelectric materials can behave quite differently. Their permittivity may change sharply near a phase transition, so materials such as barium titanate and related compositions are normally characterized over a range of temperatures rather than treating temperature testing as an optional extra.

The practical outcome is that the dissipation factor at 400 °C can exceed the room-temperature value by more than an order of magnitude in a material that looks excellent on a datasheet.

Apparatus and Electrodes

Fired-on electrodes are essential, not preferable, here. A pressed contact or a foil-and-grease arrangement will not survive 500 °C, and any air gap that remains at temperature introduces the same series-capacitance error that afflicts ambient measurement, made worse by differential expansion between specimen and electrode.
Electrode material has to suit the temperature, and silver migrates and can degrade across part of the range. Platinum is the usual choice where the full 500 °C is needed.

Where the Measurement Goes Wrong

Temperature gradient across the specimen is the first risk. If the thermocouple is reading the furnace air instead of the specimen, the reported temperature may not be the temperature the material has reached. That puts every property on the curve at the wrong point on the temperature axis. The problem is that the data itself may give no indication this happened.
The second risk is the electrode coming loose at high temperature. As the specimen expands, a fired-on electrode can lift or blister and lose contact partway through the test; on the plot, that can look a lot like a real material transition. It isn’t. The electrode failed, and the ceramic is fine.

Ceramics with alkali in them have one more issue. As temperature rises, more current flows along the part’s surface. Without a guard electrode, the instrument picks that up and adds it to the dielectric loss reading. So the dissipation factor you get back is mostly a surface value. It won’t tell you much about the bulk ceramic, which is usually what you need to know.
Absorbed moisture matters at the start of a run. Porous ceramics hold water that drives off during the first heating, so a first-run curve differs from every subsequent one, and a pre-dry or conditioning cycle is usually needed.

Specimen Requirements

We need flat discs or plates. Both faces should be parallel, and the thickness should stay the same from one side to the other. Cut them a little wider than the guarded electrode. A small border all the way round keeps the electrode away from the edge.

We take thickness measurements carefully because the dielectric constant is calculated from that figure. Any error there carries straight into your result. If you have porosity or bulk density data, include it with the samples. Both can move permittivity a fair amount. We often see customers test two lots, get two different readings, and then find the lots simply had different densities.

Let us know what the ceramic is made of. A rough idea is enough. Things like alkali content, the amount of glassy phase, or whether it contains a ferroelectric component can affect how readings change as temperature rises. If we have that information before the test, we can choose a sensible temperature range and step size for your parts.

And send a few extras. Sometimes a part cracks during electrode firing or thermal cycling, and it’s better to have a backup than to wait for a new batch.

Reporting

Permittivity and dissipation factor at every frequency and temperature point, supplied as tables and as curves against both variables.
With your results, you also get the full test conditions. That means the electrode material and how it was fired, the guard setup, heating rate, how long we held each temperature, the atmosphere, specimen thickness, and the bulk density we measured.

We leave the data as-is. No smoothing. When we see something odd, we write it down and explain it. For example, a loss peak might move across the frequency range as the part heats up. You might see a jump in permittivity close to a phase transition. Sometimes the heating curve and cooling curve don’t line up. All of this points to how your ceramic actually behaves, so we flag it in the report instead of burying it in a table.