It is at least as high as several hundrends of GHz, so above microwave frequencies into infrared.

I guess it is limited by plasmonic effects and losses in metals used, e.g.

https://www.mwrf.com/technologies/passive-components/article/21846960/dielectric-microstrip-lines-achieve-low-loss-at-thz-frequencies

Thanks, Nikolay Pavlov, for your response.

I tend to push microstrip lines a bit higher in frequency but the comments here are very useful. Much like coax, operating frequency can be pushed higher with smaller cross-sections. Unfortunately, insertion loss increases and power handling decreases.

With microstrip, I have often been curious about the best impedance for lowest loss. My suspicion is that it is probably in the vicinity of 70 ohms but I have never looked too deeply to make that case.

Thank you, Rilwan Usman for your contribution.

I was thinking about your response. You said: Unfortunately, insertion loss increases and power handling decreases. Is the increase of insertion loss due to the radiation effects? What do you think?

Best regards,

Fabrício Simoes

That's what I think. Read the following write up:

Microstrip is one of the most popular microwave transmission-line technologies. Ideally, it would operate without loss, and without adding noise or degrading signal integrity. In the real world, however, microstrip transmission lines suffer from three types of losses: conductor, dielectric, and radiation losses. The first two are fairly well understood, with conductor loss dependent on the type of metal used for the transmission lines and the dimensions of those lines (with loss increasing with length) and dielectric loss a function of the type of printed-circuit-board (PCB) substrate material. Radiation losses can stem from numerous factors, including the type of dielectric material, its thickness, and the shapes of transmission-line structures in a microstrip circuit. Radiation loss also depends very much on frequency, increasing with increasing frequency.

Radiation loss is not exclusive to microstrip, although its effects are more noticeable in microstrip than in stripline or coplanar-waveguide circuits. Stripline transmission lines, because they are enclosed by dielectric material, exhibit almost no radiation loss. Coplanar-waveguide circuits suffer minimal radiation losses, except at transmission-line discontinuities, such as a junction or step. But in certain microstrip circuits, such as filters and couplers, radiation loss can significantly impact performance, resulting in increased insertion loss, changes in signal phase, and shifts in resonant frequency. Understanding more about microstrip radiation losses and why they occur can help to control those losses.

That ideal microstrip transmission line mentioned earlier was without junctions that could increase radiation losses. Real microstrip circuits have transmission lines with their share of discontinuities, such as steps or changes in width, bends, open ends, gaps, and junctions. These discontinuities are typically an integral part of a high-frequency circuit design. For example, microstrip meander lines are slow-wave structures that are typically used to realize a shift in phase. Unfortunately, every deviation from a straight, continuous microstrip transmission line invites radiation of energy away from the metal conductor and dielectric material, effectively dissipating energy from the microstrip circuit into free space. In addition, the close proximity of conductive lines in a meander line configuration can result in the coupling of energy from adjacent lines, also causing losses.

Many circuit designers have learned to limit the number of discontinuities in a microstrip circuit in order to minimize radiation losses. And when discontinuities are necessary, creative designers seek out forms that will not invite radiation effects. The shape of a transmission-line discontinuity can influence the amount of radiation loss, with sharper edges on junctions causing higher radiation losses than rounded edges or more gradual transitions in step junctions, as used for impedance matching in low-noise amplifiers.

Microstrip transmission lines with excessive radiation loss can be thought of as antennas, albeit unintended ones. Even in a microstrip antenna or antenna array, unwanted radiation losses can occur at junctions and coupling points, resulting in degraded antenna efficiency, especially at higher frequencies. Microstrip radiation losses can be controlled, but it requires careful choices in dielectric substrate, the thickness of that substrate, and the geometry of the microstrip circuitry.

One of those choices is in the relative dielectric constant for the PCB material. Dielectric substrates with low dielectric constants, such as less than 5.0, have often been used in microwave microstrip circuits because of their low cost and versatility: they yield larger circuit geometries for a given frequency compared to circuit materials with higher dielectric constants, helping to simplify fabrication at higher frequencies. With the lower dielectric constant, however, less of the conducted electromagnetic (EM) energy is concentrated in the substrate and the microstrip metal conductor, leading to losses as a result of radiation effects.

Switching to a higher-dielectric-constant circuit material will reduce radiation losses because a greater part of the EM field is concentrated in the dielectric material between the ground plane and the microstrip metal conductor. Of course, the increase in substrate dielectric constant will also miniaturize the dimensions of the microstrip circuitry, roughly decreasing in size by a factor equal to the square root of the dielectric constant. If making such a switch in choice of dielectric material, attention should also be paid to other material parameters, so as to not trade lower radiation loss for higher dielectric loss.

The thickness of a PCB material also impacts the amount of radiation loss exhibited by a microstrip circuit at higher frequencies. For lower-dielectric-constant circuit materials, radiation losses can be reduced by using a thinner dielectric substrate. Radiation losses can also be reduced with a thicker substrate material having a higher dielectric constant, although a further reduction in radiation losses can also be achieved by using a thinner material at that higher dielectric constant. The shrinking of circuit dimensions that comes from using a substrate with higher dielectric constant can be a concern at higher frequencies, such as for millimeter-wave circuits, where the circuit dimensions and tolerances may become critical for some fabrication processes.

An example of a commercial PCB laminate available at higher dielectric-constant value is RO3010™ laminate from Rogers Corp. (www.rogerscorp.com). It is a ceramic-filled PTFE composite that can be specified with dielectric-constant values of 3.00, 6.15, and 10.2, each measured in the z direction at 10 GHz. Although the substrates in the product family exhibit different dielectric constants, they offer consistent mechanical properties, allowing designers to mix dielectric-constant values in multilayer circuit-board assemblies. The high-dielectric-constant material is available in different thicknesses, from 0.005 in. (0.13 mm) to 0.050 in. (1.28 mm), so that designers can also use that circuit-material parameter when combating the effects of radiation loss. And the material sacrifices nothing in terms of dielectric loss, with low dissipation factor of 0.0022 at 10 GHz, which is on a par with lower-dielectric-constant circuit materials.

Similarly, the TMM® circuit laminates from Rogers Corp. (www.rogerscorp.com) are available with high dielectric constant values such as 9.2 and 9.8, both measured in the z direction at 10 GHz. These are ceramic thermoset polymer composite materials with excellent mechanical and electrical stability, (evidenced by its exceptionally low thermal coefficient of dielectric constant)but also with low dielectric loss, as evidenced by dissipation factor of 0.0023 or better at 10 GHz for all TMM materials.

Radiation loss can be a concern in microstrip circuits at higher microwave frequencies. Fortunately, its effect can be minimized by reducing the number of discontinuities in a circuit design, and by carefully weighing the choice of circuit substrate material in terms of thickness and dielectric constant, where thinner materials and higher dielectric constants can both contribute to lower radiation losses.

Best Regard.

Rilwan Usman.

Dear Yongle Wu

Best regards,

Fabrício Simoes

Thanks, Rilwan Usman

Best regards,

Fabricio Simoes

Dear Fabricio,

Microstrip lines can be used over a wide frequency range. The questions you have to ask yourself is, up to which frequency it does make sense for you. The first "limitation" is the substrate thickness. This parameter should be sufficiently below lambda / 4 considering the dielectric constant of the laminate. Hence, microstrip lines designed for high frequencies show lower substrate thickness and also a lower width of the copper trace (for a constant characteristic impedance). The lower width of the copper trace results in a higher loss of the microstrip line. Therefore, the trade-off between frequency range and losses must be considered. This trade-off limits the application of microstrip lines especially at frequencies above 70 - 100 GHz where alternative types of waveguides can be implemented with reasonable effort.

Best regards,

Michael

Dear Michael Gadringer , thank you for your contribution.

Best regards,

Fabricio Simoes

I agree with Michael. The performance of microstrip lines can be used over a wide frequency range. The upper limit is the frequency at which significant coupling occurs between the quasi-TEM mode (fundamental mode of microstrip line) and lowest order surface wave spurious mode. Therefore, it is suggested to operate below the cut-off frequency of the first higher-order mode. For example, the maximum usable frequency of the microstrip line structure using a quartz substrate (0.5 mm of thickness) is around 100 GHz. In addition, you may care about your microstrip line can work as a leaky-wave antenna (if the line width is too wide or the leaky mode is accidentally excited)

Dear Huy Nam Chu , thanks you for your response.

Best regards,

Fabricio Simoes

94 GHz radar and communication antenna use microstrip patch antenna arrays.

You can see this lecture

https://drive.google.com/file/d/1EsbJJMM_PqDN2HKMdqshxCX7-nTljAeD/view?usp=sharing

Vadym Slyusar asked me to add this

Microstrip might be ok because it will be very short. Metal waveguide works well. Dielectric waveguide would be an option.

http://web.gps.caltech.edu/~gab/spectrometers/Ribbon-guide-Applied-Optics.pdf

Thank you, Malcolm White ! Your web-link gives the nice information that Microstrip lines can be used on 10 THz as well.

The are physical frequency limit of the microstrip lines can be set by the substrate properties and the metal line dimensions.

There are the dielectric losses of the substrate. There is also the resistive losses by the transmission line conductance. These cause attenuation of the transmitted signal that can not be tolerated in the system.

There are also geometrical limits set by the possible dimensions of the different features such as the width of the transmission line metal and its thickness in addition to the substrate thickness.

As said by Malcolm, the waveguides are more suitable for high frequency rather than conventional transmission lines.

Best wishes

Dear Abdelhalim abdelnaby Zekry , what about transmitted lines on the basis of T-type metamaterials?

I can not say anything about a maximum limit but i can say something about an efficient frequency range for microstrip low-pass filter. It is about 0-4 GHz. If you want to work over 4 GHz, you may not determine S-parameters..

Thanks Kemal Güvenli for your interest. But, I would like to talk about your answer. Why is it not possible to determine the S-parameters ? What are the reasons ?

Kemal Güvenli is mistaken. Microstrip S-parameters can be measured up to 40 GHz and beyond. You may need to create your own standards for calibration, of a VNA, for instance.

Malcolm White Fabrício Gerônimo Simões Silva Microstrip structures become unstable at high frequency, especially on FR4 structures (fc > 4 GHz). Therefore, you may not get the desired S-parameter results (S11 and S21). My advice is to choose SIW structures for high frequency regions, it can be very beneficial for your design. This is not related to measuring device, soldering, etc. sensitivities.

There are good quality substrates that can be used well above 4 GHz. FR4 is cheap but there are plenty of higher quality substrates that work, including ceramic. 10 GHz microstrip antennas work well.