Microstrip, Printed, and Integrated Antennas

Introduction and context

Microstrip and printed antennas form the electromagnetic interface between modern wireless electronics and free space. Unlike conventional wire or waveguide antennas, they are fabricated on dielectric laminates, semiconductor packages, flexible films, or multilayer circuit boards. Their low profile, low mass, compatibility with photolithography and printed-circuit manufacturing, and ability to integrate with transceivers make them central to WLAN, GNSS, cellular, radar, satellite, automotive, biomedical, and Internet-of-Things systems.

The same properties that make these antennas attractive also create difficult design compromises. A thin substrate can reduce profile but narrow bandwidth; a high-permittivity material can miniaturize the radiator but reduce efficiency; an integrated feed can reduce interconnect loss but increase electromagnetic coupling to the package and electronics. Consequently, antenna design at postgraduate level is not merely the selection of a patch dimension. It is the controlled synthesis of radiator, substrate, feed, ground plane, enclosure, excitation, fabrication process, and electromagnetic environment.

Learning objectives

After studying this chapter, the reader should be able to:

1. Explain the field mechanism and equivalent models of printed antennas.
2. Design a rectangular microstrip patch for a specified resonant frequency.
3. Distinguish microstrip, printed monopole, slot, planar inverted-F, substrate-integrated, and chip/package antennas.
4. Analyse bandwidth, impedance, efficiency, polarization, surface waves, and size-reduction techniques.
5. Select suitable feeding, substrate, simulation, and measurement methods.
6. Evaluate practical trade-offs in an integrated wireless product.

1. Printed-antenna foundations

A conventional microstrip patch consists of a conducting patch on one side of a dielectric substrate and a conducting ground plane on the other. The substrate has relative permittivity εᵣ, thickness h, and loss tangent tan δ. The patch is usually much thinner than a wavelength, so its thickness t is often treated as negligible in first-order analysis. Radiation occurs mainly from fringing electric fields at discontinuities, especially at open patch edges.

The most widely used first-order model is the transmission-line model. A rectangular patch excited in its fundamental mode behaves approximately as two radiating slots separated by a resonant section. The slots are in phase for the dominant TM₁₀ mode, producing broadside radiation. The patch length is approximately half a guided wavelength, while the width strongly influences radiation conductance, impedance, and bandwidth.

Important terminology includes:

• Broadside radiation: maximum radiation normal to the antenna plane.
• End-fire radiation: maximum radiation approximately parallel to the plane.
• Surface wave: a wave bound mainly to the dielectric interface and ground structure.
• Effective permittivity: an equivalent permittivity representing fields partly in dielectric and partly in air.
• Fractional bandwidth: bandwidth divided by centre frequency, commonly expressed as a percentage.
• Realized gain: gain including mismatch loss.
• Specific absorption rate (SAR): absorbed electromagnetic power per unit mass in biological tissue.

1.1 Dominant patch mode

For a rectangular patch whose length is along x and width along y, the TM₁₀ mode has one half-cycle variation along length and negligible variation along width. Its approximate resonant frequency is

where c is the speed of light in free space, L_eff is the effective electrical length, and ε_eff is effective relative permittivity. The approximation assumes a thin, homogeneous substrate, a sufficiently large ground plane, and operation near the fundamental mode.

The physical length is shorter than L_eff because the fringing fields extend beyond both radiating edges:

where L is the physical patch length and ∆L is the length extension at one radiating edge.

For a common quasi-static design approximation,

where W is patch width. A frequently used edge-extension approximation is

These expressions are design estimates, not substitutes for full-wave simulation. They become less reliable for very thick substrates, high permittivity, electrically small grounds, strong coupling, or multilayer structures.

A conventional width estimate is

where f₀ is the desired resonant frequency. The width is commonly selected near this value because it provides useful radiation conductance and reasonable input impedance. It is not a universal optimum.

1.2 Equivalent circuit interpretation

Each radiating edge can be represented approximately by a slot conductance G and susceptance B. Two slots, the patch transmission line, and the feed form a resonant network. At resonance, the input reactance is near zero, but the input resistance depends strongly on feed position and geometry. A coaxial probe, inset microstrip feed, aperture coupling, or proximity coupling may therefore be used to obtain a practical 50 Ω match.

The input reflection coefficient is

where Z_in is antenna input impedance and Z₀ is the reference impedance, usually 50 Ω. The voltage standing-wave ratio is

for |Γ| < 1. The return loss is often reported as −20 log₁₀|Γ| in decibels. These quantities describe mismatch, not radiation efficiency.

2. A first-order rectangular-patch design

Consider a patch designed for f₀ = 2.45 GHz on a substrate with εᵣ = 4.4 and h = 1.6 mm. Assume copper thickness is small relative to the dimensions and use the dominant TM₁₀ approximation.

First, the free-space wavelength is

The width estimate is

Using W/h ≈ 20.75 gives ε_eff ≈ 4.08. The edge-extension expression then gives approximately ∆L ≈ 0.74 mm. Thus,

and

An initial design is therefore W ≈ 33.2 mm and L ≈ 28.8 mm. In practice, the final dimensions are tuned using a calibrated full-wave model and prototype measurements. The connector, finite ground plane, solder mask, dielectric tolerance, and nearby enclosure can shift the resonance significantly.

3. Feeding techniques and their design implications

Microstrip-line feed

A microstrip feed is coplanar with the patch and is easy to fabricate. An inset feed places the feed point inside the patch, where the input resistance is usually lower than at the edge. The inset depth is adjusted to obtain the desired resistance, but the notch introduces capacitance and must be included in simulation.

Coaxial probe feed

A probe connects the patch to a connector or transmission line through the substrate. It permits flexible feed placement and can provide good matching. However, probe inductance increases with substrate thickness and may limit bandwidth, especially at millimetre-wave frequencies.

Aperture coupling

A slot in the ground plane couples energy from a feed line beneath the ground to the patch above. The feed and radiator can be separated, which improves isolation and permits independent optimization. The structure is multilayer and can suffer from unwanted slot radiation and fabrication alignment sensitivity.

Proximity coupling

A feed line is electromagnetically coupled to the patch without a direct electrical connection. This can provide broad bandwidth, but alignment and layer registration are important.

Integrated and differential feeds

Modern radio chips often provide differential outputs. A balanced slot, dipole, or folded radiator may be preferable to a single-ended patch. If a differential port drives an unbalanced structure, common-mode currents can flow on the package or cable, changing radiation and increasing electromagnetic interference. Baluns, transformers, Marchand structures, or carefully designed common-mode chokes may be required.

4. Printed antenna families

Printed monopoles and dipoles

A printed monopole uses a planar conductor over a ground plane and commonly provides wider impedance bandwidth than a narrow rectangular patch. Tapered, elliptical, circular, and meandered shapes are used to control current paths. A printed dipole is usually balanced and may be fed through a balun or differential line.

Planar inverted-F antenna

The planar inverted-F antenna (PIFA) is a shorted patch or plate above a ground plane, fed at a position chosen for matching. The shorting wall or strip reduces the resonant path compared with a quarter-wave monopole. PIFAs are common in handsets because they can be folded into a small volume and can support multiple resonances.

Slot antennas

A slot cut in a conducting plane is complementary to a dipole in several electromagnetic respects. Slot length, width, surrounding ground, and feed position determine resonance and coupling. Slots are useful where the radiating aperture should be integrated into a metal chassis or ground plane.

Dielectric resonator and substrate-integrated antennas

A dielectric resonator antenna radiates through resonant modes of a high-permittivity dielectric body. A substrate-integrated waveguide antenna uses rows of metallized vias to emulate waveguide sidewalls within a planar substrate. SIW structures are attractive at microwave and millimetre-wave frequencies because they combine low-loss guided-wave behaviour with planar manufacturing, although via spacing, leakage, and conductor loss must be controlled.

5. Performance metrics and physical trade-offs

Antenna radiation efficiency is

where P_rad is radiated power and P_acc is accepted power. Total efficiency includes mismatch:

The realized gain is

where D is directivity. Gain and directivity are dimensionless ratios; decibel values are obtained with 10 log₁₀ of the ratio.

A useful bandwidth definition must specify a criterion. For example, impedance bandwidth may be the frequency range for which |S₁₁| ≤ −10 dB. Fractional bandwidth is

where f_L and f_H are lower and upper band limits and f_c is the chosen centre frequency, often (f_H + f_L)/2 for a narrow symmetric band.

The principal trade-offs are:

Surface waves and finite grounds

A dielectric substrate can support surface waves when its electrical thickness is significant. Surface waves reduce useful radiation, distort patterns, increase mutual coupling, and can produce scan blindness in arrays. A finite ground plane also permits currents to flow around its edges, so measured patterns may differ substantially from infinite-ground simulations. In compact devices, the chassis is part of the antenna system rather than an incidental boundary.

Polarization and diversity

A single rectangular patch generally produces linear polarization. Circular polarization can be obtained by exciting two orthogonal modes with equal amplitude and 90° phase difference. Truncated corners, diagonal slots, perturbation strips, or dual feeds are common methods. Axial ratio, not merely return loss, must be evaluated across the required angular region.

Multiple-input multiple-output systems require low envelope correlation, adequate efficiency, and acceptable isolation. Pattern diversity, polarization diversity, spatial separation, neutralization structures, and decoupling networks may be combined. Isolation alone is insufficient: two antennas can have low S₂₁ yet still show poor diversity because of correlated radiation patterns.

6. Miniaturization and bandwidth enhancement

Miniaturization increases the ratio of stored reactive energy to radiated energy. This generally lowers radiation efficiency and bandwidth, a consequence related to fundamental small-antenna bounds. Practical methods include meandering, shorting pins, high-permittivity substrates, folded radiators, capacitive loading, and dielectric or magnetic loading.

Bandwidth can be increased by stacking patches, adding parasitic elements, using thick low-loss substrates, employing aperture or proximity coupling, introducing multiple resonant paths, or combining distinct radiator modes. Multiresonant designs should not be judged only by the union of |S₁₁| curves; gain, efficiency, pattern stability, group delay, and polarization must also remain acceptable.

At wide bandwidth, the antenna and feed are distributed structures. Lumped equivalent circuits are useful for interpretation but cannot fully represent higher-order modes, surface waves, radiation from feed lines, or enclosure coupling.

7. Integrated, package, and on-chip antennas

An integrated antenna is designed as part of a larger electronic assembly. Categories include:

1. Board-level antennas, printed on a PCB with the radio module.
2. Package antennas, embedded in or mounted on an IC package, laminate, or lead frame.
3. Antenna-in-package (AiP) systems, in which antenna arrays and RF front ends share a package.
4. On-chip antennas, fabricated in CMOS, SiGe, GaAs, or another semiconductor process.

On-chip antennas face low-resistivity silicon loss, thin dielectric stacks, metal loss, limited area, and interaction with passivation and interconnect layers. Their efficiency may be modest, but they can be valuable at millimetre-wave and terahertz frequencies because the wavelength is small and interconnect parasitics are significant. Backside removal, high-resistivity substrates, patterned ground shields, top-metal radiators, and wafer-level packaging can improve performance.

At 5G, 6G, radar, and satellite frequencies, AiP enables short RF paths and dense phased arrays. The design must include thermal structures, mechanical tolerances, radome effects, calibration, electromagnetic compatibility, and semiconductor noise coupling. Antenna efficiency, beam-steering range, scan loss, sidelobe level, effective isotropic radiated power, and receiver sensitivity must be evaluated together.

8. Simulation, fabrication, and measurement workflow

A defensible design workflow is as follows:

1. Convert system requirements into frequency bands, polarization, gain, bandwidth, allowed volume, efficiency, power, and regulatory limits.
2. Select substrate, conductor, stack-up, and manufacturing tolerances.
3. Establish a first-order analytical design.
4. Build a three-dimensional electromagnetic model including finite ground, connector or chip port, enclosure, radome, battery, and nearby metals.
5. Perform parameter sweeps or optimization, while constraining manufacturability.
6. Fabricate a coupon or prototype with the intended materials and process.
7. Calibrate the vector network analyser using an appropriate SOLT, TRL, or other method at the measurement reference plane.
8. Measure S-parameters, efficiency, gain, radiation patterns, polarization, and environmental sensitivity.
9. Correlate simulation and measurement, identify model errors, and update the design.

Mesh convergence, port definition, material dispersion, conductor roughness, solder mask, and connector modelling are frequent sources of error. A good return-loss agreement does not prove correct gain or pattern prediction. Anechoic chambers, reverberation chambers, Wheeler-cap measurements, and calibrated gain-comparison methods each have different strengths and limitations.

9. Applied case: illustrative compact dual-band terminal antenna

Background

Consider an illustrative handheld terminal requiring operation in two bands near 2.4 GHz and 5.8 GHz. The available clearance above a metal battery shield is only 8 mm, and the antenna must support two orthogonal linear polarizations for diversity.

Problem

A single conventional patch would require excessive area and would not naturally provide two well-separated bands. The metal shield and display frame also act as parasitic conductors, while the user’s hand detunes the lower band.

Analysis

A PIFA-like radiator is selected because its shorting strip reduces the required resonant path. A longer current path is assigned to the lower band and a slot or branch is introduced for the upper band. Two feeds are placed to excite distinct current distributions. The design target is |S₁₁| below −10 dB over both bands, total efficiency above a project-defined threshold, and acceptable envelope correlation.

The first layout is simulated with the battery shield, display frame, plastic housing, and lossy hand phantom. Parametric variables include shorting-strip width, branch length, feed positions, ground clearance, and slot dimensions. The design is optimized using full-wave results rather than relying on an isolated-antenna model.

Approach and outcome

The final prototype is tuned with small matching components near the feed, but the matching network is kept simple to limit loss and production variation. Measurements are made in free space and beside the hand phantom. The lower resonance shifts downward under hand loading, so the radiator is intentionally tuned slightly high in the unloaded condition. Diversity performance is reported using efficiency and correlation, not only S₂₁.

Learning points

The case demonstrates that the product chassis is part of the electromagnetic design, detuning must be evaluated in realistic use conditions, and miniaturization is inseparable from efficiency and bandwidth compromises. It also shows why a successful design process combines analytical estimates, full-wave simulation, tolerance analysis, and measurement.

10. Regulation, safety, and responsible engineering

Printed antennas must satisfy national and regional spectrum regulations, including occupied bandwidth, spurious emissions, transmit power, and exposure limits. In portable and wearable devices, SAR or power-density compliance may determine the allowable radiator location and operating power. Designers should document assumptions, test configurations, uncertainty, and worst-case user conditions.

Electromagnetic compatibility is also a governance issue: an antenna can increase susceptibility of medical, automotive, or avionics systems, or unintentionally couple energy into sensitive circuits. Responsible engineering therefore includes traceable simulation models, calibrated measurements, production controls, privacy-aware wireless deployment, and honest reporting of efficiency and gain under specified conditions.

11. Contemporary developments and research directions

Current research includes reconfigurable antennas using PIN diodes, varactors, RF MEMS, liquid conductors, and tunable materials; flexible and textile antennas; metasurface and transmitarray apertures; additive manufacturing; chiplet and heterogeneous integration; and machine-learning-assisted optimization. AI can accelerate surrogate modelling, inverse design, and tolerance exploration, but its outputs require physical validation. A model trained on one substrate stack or enclosure should not be assumed valid for another.

Emerging terahertz and 6G systems require co-design of antenna, package, thermal path, semiconductor circuit, and channel. The research challenge is increasingly system-level: the best antenna is not necessarily the one with maximum isolated gain, but the one that maximizes link reliability, energy efficiency, manufacturability, calibration stability, and compliance.

Chapter Summary

Microstrip and printed antennas use planar conductors, dielectric substrates, ground planes, and fringing fields to produce compact, manufacturable radiators. The rectangular patch is understood through the transmission-line and cavity viewpoints: its effective length is approximately half a guided wavelength, with fringing fields producing a length extension. Width, substrate permittivity, thickness, feed technique, and ground size control impedance, bandwidth, efficiency, and pattern.

Printed monopoles, dipoles, PIFAs, slots, dielectric resonators, SIW antennas, package antennas, and on-chip antennas extend the concept to diverse applications. Performance must be assessed through reflection coefficient, bandwidth, efficiency, gain, polarization, correlation, and realistic environmental loading. Miniaturization commonly reduces bandwidth and efficiency, while bandwidth enhancement often requires additional layers, modes, or volume.

Professional design requires analytical estimates followed by full-wave simulation, tolerance analysis, calibrated measurement, and system-level validation. In integrated products, the chassis, package, battery, radome, user, thermal structure, and electronics are electromagnetic parts of the antenna. Regulatory compliance, exposure safety, EMC, reproducibility, and transparent reporting are essential engineering responsibilities.

Key Terms

Microstrip patch: A planar conducting radiator above a ground plane and dielectric substrate.

Printed antenna: An antenna fabricated using planar circuit-board, thin-film, or additive processes.

Fringing field: Electric field extending beyond a physical conductor edge.

Effective permittivity: Equivalent relative permittivity representing fields in dielectric and air.

TM₁₀ mode: Dominant rectangular-patch mode with one half-cycle variation along patch length.

Surface wave: Guided electromagnetic field propagating near a dielectric interface.

Inset feed: A microstrip feed recessed into a patch to control input resistance.

PIFA: Planar inverted-F antenna using a shorting element and a capacitive or planar radiator.

Slot antenna: Radiator formed by an aperture in a conducting surface.

SIW: Substrate-integrated waveguide formed using planar metal and via fences.

Radiation efficiency: Radiated power divided by accepted power.

Realized gain: Gain including impedance mismatch loss.

Fractional bandwidth: Bandwidth divided by centre frequency, expressed as a percentage.

Antenna-in-package: Antenna integrated within or directly on an electronic package.

Axial ratio: Ratio of major to minor polarization ellipse axes.

Envelope correlation coefficient: Measure of similarity between diversity-channel responses.

Surface-wave loss: Loss or performance degradation caused by power coupled into substrate modes.

SAR: Absorbed electromagnetic power per unit mass of biological tissue.

Full-wave simulation: Numerical solution of Maxwell’s equations for the complete electromagnetic structure.

Review Questions

Short Answer Questions

1. Why is a rectangular patch physical length shorter than half a guided wavelength?
2. Define effective permittivity and explain why it is less than the substrate permittivity in a typical microstrip antenna.
3. Distinguish radiation efficiency from total efficiency.
4. State two advantages and two limitations of aperture coupling.
5. Why can the chassis of a portable device not be treated as an electrically irrelevant object?

Long Answer Questions

1. Explain the transmission-line model of a rectangular microstrip patch and derive the principal first-order design equations.
2. Compare microstrip-line, coaxial-probe, aperture-coupled, and proximity-coupled feeds.
3. Discuss the physical trade-offs among substrate thickness, relative permittivity, bandwidth, surface waves, and efficiency.
4. Explain the operating principles and applications of PIFAs, printed monopoles, slot antennas, SIW antennas, and antenna-in-package systems.
5. Develop a complete simulation-to-measurement workflow for a printed antenna and identify major sources of disagreement.

Analytical / Application Questions

1. Design an initial rectangular patch at 3.5 GHz on a substrate with εᵣ = 3.0 and h = 0.8 mm. State assumptions, calculate W and L using the first-order equations, and explain how you would refine the design.
2. A compact antenna has excellent |S₁₁| but poor total efficiency when installed beside a battery. Diagnose at least four possible causes and propose measurements or simulations to separate them.
3. Select an antenna architecture for a 28 GHz five-channel phased-array module in a thermally constrained package. Justify the choice using integration, scan, loss, calibration, manufacturability, and regulatory considerations.

Multiple Choice Questions with Answers

Question 1

For the dominant TM₁₀ mode of a rectangular patch, the principal field variation is along the:

A. Patch length

B. Patch width only

C. Substrate thickness only

D. Connector cable

Correct Answer: A. Patch length

Question 2

Which quantity includes impedance mismatch as well as radiation efficiency?

A. Directivity

B. Realized gain

C. Radiation resistance

D. Beamwidth

Correct Answer: B. Realized gain

Question 3

A common first-order estimate for patch width is based primarily on:

A. c/(2f₀)√[2/(εᵣ+1)]

B. c·f₀·εᵣ

C. 2πf₀h

D. h/(c f₀)

Correct Answer: A. c/(2f₀)√[2/(εᵣ+1)]

Question 4

The principal purpose of an inset feed is to:

A. Increase conductor thickness

B. Control the patch input resistance

C. Eliminate all surface waves

D. Create a magnetic substrate

Correct Answer: B. Control the patch input resistance

Question 5

If |Γ| = 0, the VSWR is:

A. 0

B. 1

C. 2

D. Infinity

Correct Answer: B. 1

Question 6

Which technique separates the feed line from the patch using a slot in the ground plane?

A. Direct edge feed

B. Coaxial probe feed

C. Aperture coupling

D. Shorting-pin feed

Correct Answer: C. Aperture coupling

Question 7

Increasing substrate relative permittivity generally causes which combination?

A. Larger size and wider bandwidth

B. Smaller size and potentially narrower bandwidth

C. Zero stored energy and higher efficiency

D. No change in resonance

Correct Answer: B. Smaller size and potentially narrower bandwidth

Question 8

A PIFA reduces resonant path length primarily through its:

A. Shorting element

B. Infinite ground plane assumption

C. Copper oxidation

D. Anechoic chamber

Correct Answer: A. Shorting element

Question 9

The total efficiency relation η_tot = (1 − |Γ|²)η_rad assumes that:

A. Mismatch and radiation efficiency are treated as separable factors

B. The antenna has zero radiation

C. The reference impedance is infinite

D. The antenna is necessarily circularly polarized

Correct Answer: A. Mismatch and radiation efficiency are treated as separable factors

Question 10

Why must an integrated handset antenna be tested with its battery and enclosure?

A. They determine the speed of light

B. They can alter current paths, resonance, loss, and radiation

C. They remove the need for calibration

D. They guarantee ideal matching

Correct Answer: B. They can alter current paths, resonance, loss, and radiation