QUESTION:

How is the dielectric constant of the different high frequency materials affected by changes in temperature?

ANSWER:

One of the most critical properties of a high frequency material is its dielectric constant, (e_r). This property can be measured in various ways (stripline, microstrip or waveguide), at various frequencies (1 MHz, 10 GHz or 77GHz) or under various conditions (temperature or humidity). The measured e_r that one obtains can be different depending on the measurement style, frequency or condition. The lack of NIST traceable standards, makes it impossible to know the exact e_r of a material and the bias that may be found in each test condition. Most suppliers of high frequency materials report material e_r on the data sheets for a given set of conditions. As long as the same testing conditions are followed, one can understand the material e_r repeatability, lot to lot. There are applications that require the material to be exposed to a wide range of temperatures and an understanding of the material behavior is essential.

To measure e_r, Rogers uses the industry standard test IPC-TM-650 2.5.5.5 @ 10 GHz, 23°C and 50% relative humidity. A stripline end coupled straight resonator, two wavelengths long at 10 GHz, is used to measure e_r and loss tangent (tan d )from the resonant frequency and circuit Q. Fixtures for this test require a circuit pattern to be etched on a thin sample of the same type of material to be tested. A view of the circuit pattern and side view of the stripline circuit (and material to test) in the fixture are given in Figure 1. The material to be tested is etched free of any copper and placed on either side of the circuit pattern. Pressure is applied to remove air entrapment for the stripline assembly that may cause higher losses to be measured.

FIGURE 1. IPC-TM-650 2.5.5.5 Circuit Pattern and Side View of Fixture Used to Measure e_r and tan d 

Having found the resonant frequency we can then calculate the dielectric constant, knowing the length of the resonator and having obtained the fringing factor effect. The loss tangent is obtained from the circuit Q by subtracting the losses due to the conductor. The equations below provide an understanding of the test. Because this is a 4 node resonator, it can also be used to measure the properties at 2.5, 5 and 7.5 GHz.

In order to measure e_r vs temperature, this test method is modified by etching the stripline resonant circuit pattern on the material to be tested. The fixture is assembled with the material and placed in a thermal cycling chamber. Too much variation in tan d  was encountered vs. temperature, so data was not used. The materials were measured from -50°C to +150°C and at various temperatures in between. The data is the normalized with respect to a base temperature (usually 23°C) in the following way ;

e_{r Normalized} = e_rr_Temp /e_{r 23°C}

Rogers high frequency materials are generally divided in two groups, thermoplastic (PTFE based) and thermoset (hydrocarbon). PTFE is an excellent high frequency resin, but it is known to have noticeable changes in electrical and mechanical performance vs. temperature (with a transition between 17°C and 21°C ). Among the first high frequency laminates introduced to the market were RT/duroid® 5880, 6006 and 6010 materials. These materials were designed to meet the needs of RF/microwave engineers, who at that time had only PTFE/woven glass as an option. RT/duroid 5880 (PTFE/random glass) addressed the need for low loss material while RT/duroid 6006 and 6010 (PTFE/ceramic) was used when high e_r was needed. Figure 2 displays the change in e_r (normalized to 23°C ) vs. temperature (°C) for these materials.

Figure 2. e_r Normalized to 23°C vs. Temperature for RT/duroid 5880, 6006 and 6010

Continuing the development of PTFE based materials that were temperature stable as well, led to the introduction of a second type of PTFE/ceramic laminates called RT/duroid 6002 material, and later to the first of the commercial options RO3003® high frequency material. In figure 3, one can see that these materials have exceptional stability of e_r vs temperature, varying less than ±0.25% for the temperature range. Improvements to the stability of e_r were also made on the high e_r options of the commercial laminates RO3006™ and RO3010™, part of the RO3000® family of high frequency materials.

  figure 4.

Figure 3. e_r Normalized to 23°C vs Temperature for RT/duroid® 6002 and RO3003™

Figure 4. e_r Normalized to 23°C vs Temperature for RO3006 and RO3003

Thermoset high frequency materials have the electrical performance associated with PTFE-based materials but in a rigid laminate form. The big advantage to the RO4000 family, is that these materials can be processed like FR4 laminates while still being low loss temperature stable materials, as can be seen in figure 5.

Figure 5. e_r Normalized to 23°C vs Temperature for RO4003C™ and RO4350B™

Circuit modeling software enables designers to cut time from the concept stage to the first prototype build. It is a desired characteristic that the material selected be temperature stable, but not always do the materials available in this group have all the properties needed for every application. For these cases, an understanding of e_r against temperature is needed to be able to simulate the performance of the design when conditions are not in a temperature controlled environment. This enables designers to save valuable time and money by being able to predict the effect of temperature change on the design.