A microstrip line is one of the most widely used ways to route high-frequency electrical signals on a flat circuit board. It consists of a thin conducting strip separated from a wider ground plane by a slab of insulating material called a dielectric substrate. If you have ever looked at the copper traces on a Wi-Fi router board or a satellite receiver, you have likely seen microstrip lines at work. Their popularity comes down to a simple advantage: they can be manufactured using the same printed-circuit-board (PCB) processes that produce the rest of a device’s electronics, which keeps costs low and makes integration straightforward. But the physics underneath that flat copper strip are more interesting, and more demanding, than first appearances suggest.
The Basic Structure
A microstrip line has three layers. On top sits a narrow conducting strip, usually copper, whose width and thickness you can control during fabrication. Below the strip is a dielectric substrate, a sheet of insulating material with a specific thickness and electrical permittivity. Beneath the substrate lies a continuous ground plane, also copper. When a signal travels along the top strip, the electromagnetic field associated with that signal extends partly through the substrate and partly through the air above the strip. This split environment is what makes microstrip lines both useful and tricky to design.
Because the field lives partly in the substrate and partly in air, the signal does not experience the full permittivity of the substrate. Instead, designers work with an “effective” permittivity, a value somewhere between the substrate’s actual permittivity and the permittivity of free space. This effective permittivity determines how fast signals travel along the line and, together with the strip’s width and the substrate’s thickness, sets the line’s characteristic impedance. Most designs target 50 ohms, which is the standard impedance for RF and microwave systems, though 75-ohm lines appear in some video and cable-television applications.
How Signals Actually Travel Along the Strip
In a perfectly symmetric transmission line, like a coaxial cable, the electromagnetic wave traveling down the line is a pure transverse electromagnetic (TEM) wave: the electric and magnetic fields are entirely perpendicular to the direction the signal moves. Microstrip cannot support a pure TEM wave because the dielectric below the strip differs from the air above it. The mismatch means a small longitudinal field component always exists. In practice, though, the longitudinal component is small enough at moderate frequencies that engineers treat the wave as “quasi-TEM,” a nearly transverse wave that behaves like a true TEM wave for most design purposes.
Research on superconducting microstrip lines has confirmed that the quasi-TEM approximation holds well up to several gigahertz. One study of YBaCuO superconducting microstrip found strong agreement between experimental measurements and simulations based on the quasi-TEM model up to 7 GHz and across a range of operating temperatures.1Physica C: Superconductivity. Modeling of microstrip quasi-TEM superconducting transmission lines, comparison with experimental results At higher frequencies, the longitudinal field components grow larger and the quasi-TEM picture starts to break down. Designers working at millimeter-wave frequencies have to account for dispersion effects that the simple model ignores, which is one reason why alternative structures sometimes replace microstrip at the highest frequencies.
Controlling Impedance Through Geometry
One of the practical strengths of microstrip is that you can set the characteristic impedance by choosing the right combination of strip width, substrate thickness, and substrate permittivity. A wider strip lowers the impedance; a thicker substrate raises it. Higher-permittivity substrates pull the impedance down for a given geometry. This means that hitting your target impedance is a matter of selecting the right substrate and then calculating the strip width, a process handled by well-established closed-form equations and by every modern RF circuit simulator.
Where this gets interesting is at the edges. Very narrow strips produce high impedances but are harder to fabricate precisely, and their losses go up because the current crowds into a smaller cross section. Very wide strips give low impedances but can start supporting higher-order waveguide modes that interfere with the desired signal. The practical range of impedances you can reliably achieve on a standard PCB is roughly 20 to 120 ohms. Outside that range, you generally need to switch to a different transmission-line topology.
Bends, junctions, and discontinuities in the strip also affect performance. A sharp 90-degree corner reflects some signal energy back toward the source. Chamfering or rounding the corner reduces the reflection. Similarly, a step change in strip width, which you might need when transitioning between two different impedance sections, introduces a small parasitic capacitance that can be compensated by adjusting the geometry. These details seem minor, but at microwave frequencies where wavelengths shrink to millimeters, a poorly managed discontinuity can wreck a circuit’s performance.
Why Substrate Material Matters So Much
The substrate sitting between the strip and the ground plane is the single biggest factor in how well a microstrip circuit performs. Two properties dominate the decision: the dielectric constant (permittivity) and the loss tangent (a measure of how much signal energy the substrate absorbs as heat). A high dielectric constant shrinks the physical size of the circuit, because the wavelength inside the substrate is shorter. But a high loss tangent means the signal decays faster as it travels, which limits how far or how cleanly you can route a trace.
FR-4, the standard glass-epoxy laminate used in everyday electronics, is cheap and easy to work with. It performs acceptably at lower frequencies and for non-critical applications. But its relatively high loss tangent and somewhat variable permittivity make it a poor choice for demanding microwave designs. Comparative studies consistently show that FR-4 underperforms specialty substrates on every RF metric that matters. One evaluation of patch antenna performance across several substrates found that FR-4 had the lowest efficiency and directivity values of all the materials tested, along with poorer impedance matching indicated by higher return loss.2Radioelectronic and Computer Systems. Performance evaluation of inset feed microstrip patch antenna parameters with different substrate materials for 5G wireless applications
On the other end of the spectrum, PTFE-based laminates like Rogers RT/duroid 5870 and Rogers RO4003C offer low loss tangents and tightly controlled permittivity. These materials are the workhorses of satellite communication, radar, and 5G base-station design. Rogers 5880, for instance, achieved efficiency values consistently above 70% across all tested thicknesses in a comparative antenna study, peaking around 86% at certain configurations, and maintained return loss well below the standard acceptability threshold across the frequency range.3Radioelectronic and Computer Systems. Performance evaluation of inset feed microstrip patch antenna parameters with different substrate materials for 5G wireless applications Another study looking at 5G patch antenna designs confirmed that Rogers RO4003C and RT/duroid 5870 demonstrated no significant dielectric losses and delivered bandwidths of roughly 11% and 8%, respectively, while FR-4 managed only a tiny fraction of a percent.4Advances in Computer Science Research. Optimizing 5G Rectangular Patch Antenna Performance with Enhanced Combinations of FR-4, Rogers RO4003C, Alumina, RT/duroid 5870, Arlon AD350, and Teflon Substrates
That said, cost plays a real role. PTFE-based substrates can be several times more expensive than FR-4 and require different handling during fabrication. For prototyping, educational projects, or applications where performance demands are moderate, FR-4 remains a reasonable choice. A comparative analysis of sustainable dielectric substrates noted that while PTFE and Rogers RT5870 perform better in RF applications, FR-4 offers reasonable RF performance at a significantly lower price point, making it viable for cost-sensitive designs.5Indian Journal of Engineering and Materials Sciences. Sustainable dielectric substrate for microstrip patch antennas: A comparative analysis The substrate decision, in practice, usually comes down to balancing electrical performance against cost and manufacturability for the specific application.
Building RF Components from Microstrip
Microstrip is not just for carrying signals from point A to point B. Engineers shape sections of microstrip into functional components: filters, couplers, power dividers, impedance transformers, and antennas. A quarter-wavelength section of microstrip, for instance, acts as an impedance transformer, and two of them arranged in a Y shape form a Wilkinson power divider, a device that splits an incoming signal into two equal outputs with good isolation between them. Filters can be built by coupling multiple resonant microstrip sections together.
The downside of these distributed-element designs is size. A quarter-wavelength section at 1 GHz is several centimeters long, which adds up quickly in a complex circuit. Researchers have developed hybrid approaches that replace distributed microstrip sections with lumped-element equivalents to shrink the footprint. One study demonstrated a modified Wilkinson power divider that replaced the conventional quarter-wavelength branches with compact LC networks, achieving a size reduction of 100% compared to the conventional distributed design while also improving harmonic suppression and frequency selectivity.6PubMed Central. Size reduction and performance improvement of a microstrip Wilkinson power divider using a hybrid design technique These hybrid techniques have become increasingly common as devices shrink and frequency bands multiply, especially in mobile and IoT hardware.
Microstrip patch antennas deserve a mention here too. A rectangular or circular patch of copper on a grounded substrate, fed by a microstrip line, radiates electromagnetic energy and functions as an antenna. These antennas are flat, lightweight, and easy to manufacture, which is why they show up in GPS receivers, smartphones, and phased-array radar panels. Their main limitation is narrow bandwidth: a basic patch antenna might cover only a few percent of its center frequency. Techniques like stacking multiple patches, cutting slots in the patch, or using thicker substrates can widen the bandwidth, but the fundamental narrowband character is hard to escape entirely.
Defected Ground Structures
A relatively recent development in microstrip design is the defected ground structure, or DGS. The idea is simple: instead of leaving the ground plane as a continuous sheet of copper, you etch deliberate slots or patterns into it beneath the microstrip line. These defects change the current distribution in the ground plane and, in turn, modify the effective capacitance and inductance of the line. The result is a way to add filtering behavior, suppress unwanted harmonics, or improve bandwidth without adding extra components.
DGS has been adopted as a technique for improving several parameters of microwave planar circuits, including narrow bandwidth, cross-polarization, and low gain.7International Journal of Antennas and Propagation. Defected Ground Structure: Fundamentals, Analysis, and Applications in Modern Wireless Trends A dumbbell-shaped slot in the ground plane, for example, introduces a resonant behavior that rejects signals at a specific frequency, effectively creating a bandstop filter with no additional parts. More complex DGS shapes, like spirals or interdigital patterns, give designers finer control over the rejection bandwidth and depth.
The tradeoff is that cutting holes in the ground plane can cause unwanted radiation from the back side of the board and complicates shielding. In systems where electromagnetic interference is a concern, or where the circuit board sits inside a metal enclosure, back-side radiation from DGS slots needs to be managed carefully, sometimes by adding a secondary ground layer beneath the defects. Still, DGS has become a standard tool in the microstrip designer’s kit, particularly for compact filter and antenna designs where board space is at a premium.
Superconducting Microstrip Lines
At cryogenic temperatures, superconducting materials carry current with essentially zero resistance, which makes them attractive for microstrip lines in applications where extremely low signal loss matters. Quantum computing is the headline example: superconducting qubits communicate through microwave-frequency signals routed along transmission lines, and any loss in those lines degrades the fragile quantum information. Standard copper microstrip at room temperature would introduce far too much noise and attenuation for these circuits to function.
Researchers have demonstrated superconducting microstrip lines built from niobium deposited on flexible polyimide substrates. These lines achieved insertion losses of less than about 0.1 dB over a 50 mm length at frequencies up to at least 14 GHz, a level of performance that copper traces cannot approach.8Superconductor Science and Technology. Flexible superconducting Nb transmission lines on thin film polyimide for quantum computing applications The flexibility of the polyimide substrate is significant because it allows the lines to be routed in three dimensions and connected between different components inside a cryogenic package without rigid connectors that introduce their own losses and reflections.
The catch, of course, is the operating environment. Niobium superconducts only below about 9 kelvin, which means the entire circuit must live inside a dilution refrigerator or similar cryogenic system. The complexity and cost of maintaining cryogenic temperatures limit superconducting microstrip to specialized applications. But for quantum processors, where even tiny amounts of loss or noise can decohere a qubit, the engineering effort is justified. As quantum computing hardware scales up and the number of signal lines per processor grows, the design of these superconducting interconnects becomes an increasingly active area of research.
Microstrip Versus Other Planar Transmission Lines
Microstrip is not the only option for routing signals on a flat board. Coplanar waveguide (CPW) places the signal strip and ground conductors on the same surface, eliminating the need for a via to connect to a ground plane on the opposite side of the board. This makes CPW easier to probe during testing and better suited for mounting surface-mount components that need a ground connection right next to the signal trace. CPW also tends to have lower dispersion at very high frequencies.
Stripline, another alternative, sandwiches the signal trace between two ground planes with dielectric filling the gaps above and below. Because the entire field is contained within the dielectric, stripline supports a true TEM mode and has no radiation loss. The cost is that it requires a multilayer board and is harder to access for testing or component mounting. Stripline is common inside multilayer PCBs where signal isolation between layers matters, such as in high-speed digital backplanes.
Microstrip’s main advantages over these alternatives are simplicity and accessibility. The signal trace sits on the top surface where you can solder components directly to it, probe it with test equipment, and visually inspect it. It requires only a two-layer board at minimum. Its radiation, while a disadvantage when you want a pure transmission line, becomes a feature when you want an antenna. And the vast library of design equations, simulation models, and fabrication guidelines accumulated over more than half a century makes microstrip the default starting point for most planar RF designs.
Losses and How Designers Manage Them
Every microstrip line loses some signal energy, and understanding where those losses come from helps explain many of the design decisions already discussed. Three mechanisms are responsible. Conductor loss comes from the finite resistivity of the copper strip and ground plane: current flowing through a real metal generates heat. This loss increases with frequency because of the skin effect, which forces current into a thinner and thinner layer at the conductor’s surface, effectively raising the resistance. Dielectric loss comes from the substrate absorbing electromagnetic energy. This is where the loss tangent of the substrate material matters most. Radiation loss comes from the open structure of the microstrip line: unlike a coaxial cable or stripline, the fields are not fully contained, and some energy leaks away as radiation.
At low frequencies, conductor loss usually dominates. At higher frequencies, dielectric loss catches up and can overtake it, especially on lossy substrates like FR-4. Radiation loss becomes significant at discontinuities such as bends, open ends, and junctions. The relative importance of these three mechanisms depends on the specific design, but the general strategy is the same everywhere: use the lowest-loss substrate you can afford, keep traces short, minimize sharp discontinuities, and choose a strip geometry that balances impedance requirements against loss.
For high-power applications, thermal management adds another layer of concern. The heat generated by conductor and dielectric losses must be dissipated before it damages the substrate or shifts the electrical properties of the line. Substrates with higher thermal conductivity, such as alumina (a ceramic), handle power better than PTFE-based laminates, which are relatively poor thermal conductors. This is why high-power microwave amplifiers and transmitters sometimes use alumina or aluminum nitride substrates despite their higher dielectric constants and more difficult machining requirements.
Fiber Weave and Other Manufacturing Realities
Most PCB substrates are not homogeneous materials. FR-4, for instance, is made of woven glass fibers impregnated with epoxy resin. The glass fibers have a different permittivity than the resin filling the gaps between them, which means the effective permittivity under a microstrip trace depends on where the trace lands relative to the weave pattern. If two parallel traces happen to sit over different parts of the weave, their signals can travel at slightly different speeds, introducing skew. In high-speed digital systems where timing margins are tight, this fiber weave effect can cause real problems.
Designers mitigate the issue by rotating the trace direction relative to the weave angle, by using substrates with tighter weave patterns that average out the inhomogeneity, or by specifying spread-glass fabrics where the fibers are more evenly distributed. These are manufacturing-level decisions that do not show up in textbook microstrip theory but matter enormously in practice. A microstrip line that looks perfect in simulation can misbehave on a real board if the fiber weave effect is ignored.
Other manufacturing concerns include etch tolerances, which determine how accurately the strip width matches the design intent, and copper surface roughness, which increases conductor loss at high frequencies by forcing current to follow a longer path along the rough surface. Premium substrates aimed at millimeter-wave applications specify both tighter etch tolerances and smoother copper foils. These process details are unglamorous compared to the electromagnetic theory, but they are often what separates a circuit that works from one that does not.