
What Is a Reflectarray Antenna?
A reflectarray antenna is a flat aperture that mimics a parabolic dish by printing an array of resonant elements—typically microstrip patches—on a grounded dielectric substrate, with each element tuned to impose a specific phase shift on the incident wave from a feed horn. The combined re-radiated field forms a collimated, high-gain beam. Gains of 25–40 dBi are routine on apertures from 10 λ to 50 λ across, at a fraction of the volume and mass of a shaped reflector. Because the entire reflecting surface is a standard PCB, the reflectarray is one of the most manufacturing-friendly high-gain antenna architectures available.
How a Reflectarray Works: Phase Compensation on a Flat Surface

A parabolic dish achieves in-phase addition at the aperture by geometry: every ray path from focus to aperture plane is the same electrical length. A flat reflectarray cannot rely on curvature, so each element must independently compensate the differential path length from the feed. The required phase at element (m, n) is:
φ(m,n) = k₀ · (|r_mn − r_f| − (r_mn − r_f) · û_beam)
where k₀ is the free-space wavenumber, r_mn is the element position vector, r_f is the feed position, and û_beam is the desired beam direction. The designer maps this continuous phase distribution onto a discrete set of element geometries, each providing a known reflection phase. A full 360° phase range is needed for an efficient design; a range below about 300° causes quantization loss and raised sidelobes.
Element Types and Phase-Tuning Methods
- Variable-size patches: The simplest approach. Patch length (or diameter, for circular patches) is swept to cover the phase range. A square patch on a grounded Rogers RO4350B substrate (Dk 3.48, 20 mil / 0.508 mm thick) at 12 GHz can sweep roughly 330° of reflection phase by varying the side length from 5.5 mm to 9.5 mm.
- Rotation technique (Huang rotation): Circularly polarized reflectarrays rotate identical elements by half the required phase shift. This gives an inherent 360° range and relaxes fabrication tolerances, since every element is the same shape.
- Multi-resonance / stacked patches: Two or three patch layers separated by bonding films widen the bandwidth by creating multiple coupled resonances. A dual-layer design can push 1-dB gain bandwidth from ~5 % to 12–15 %.
- Delay-line elements: A patch connected to a stub or meandered line whose length sets the phase. Useful when a linear phase–length relationship is preferred for broadband compensation.
- Sub-wavelength elements: Smaller-than-λ/2 unit cells (typically 0.3 λ–0.4 λ) reduce grating lobes when the beam is scanned off broadside, at the cost of tighter etching tolerances.
Understanding the relationship between antenna gain, directivity and efficiency is essential when budgeting how much of the theoretical aperture gain the real reflectarray will deliver.
Gain Budget and Aperture Efficiency
The directivity of an ideal uniformly illuminated aperture of area A is D = 4πA / λ². Real reflectarray gain falls below this by several loss mechanisms:
| Loss Mechanism | Typical Impact | Mitigation |
|---|---|---|
| Feed spillover | 0.5–1.5 dB | Optimize f/D ratio (typically 0.8–1.0 for center-fed) |
| Feed illumination taper | 1–2 dB | Trade off against spillover; –10 dB edge taper is a common compromise |
| Phase quantization | 0.2–0.8 dB | Finer element sizing resolution; use ≥ 8 discrete phase states |
| Element loss (dielectric + conductor) | 0.3–1.0 dB | Low-loss laminate, 1 oz or heavier copper, ENIG or immersion silver finish |
| Feed blockage (center-fed) | 0.2–0.5 dB | Offset feed eliminates blockage at the cost of asymmetric illumination |
| Phase error from Dk tolerance | 0.1–0.4 dB | Specify laminate Dk tolerance ≤ ±0.02; calibrate element library with measured Dk |
Combining these, a well-designed single-layer reflectarray achieves 50–60 % aperture efficiency. A 30 cm × 30 cm aperture at 12.5 GHz (Ku-band) therefore delivers roughly 28–30 dBi of realized gain. For a deeper walkthrough of how aperture area converts to gain, see antenna gain calculation methods.
Substrate and Material Selection
The substrate under the patch elements sets three things at once: the resonant frequency of each element, the achievable phase range, and the ohmic loss per reflection. Choosing the right high-frequency laminate is the single most consequential material decision in a reflectarray.
Material Comparison for Reflectarray Substrates
| Laminate | Dk | Df (10 GHz) | Typical Use Case | Relative Cost |
|---|---|---|---|---|
| Rogers RT/duroid 5880 | 2.20 | 0.0009 | Low-loss Ku/Ka-band arrays | High |
| Rogers RO3003G2 | 3.00 | 0.0010 | Compact elements, good Dk stability | Medium–High |
| Rogers RO4350B | 3.48 | 0.0037 | Cost-effective prototyping up to ~20 GHz | Medium |
| Rogers RO4360G2 | 6.15 | 0.0038 | Miniaturized elements, reduced inter-element spacing | Medium |
| Isola Astra MT77 | 3.00 | 0.0017 | Multi-layer stacked designs, good processability | Medium |
Choose low-Dk (2.2–3.0) laminates when loss is the priority and the aperture can accommodate larger elements. Choose higher-Dk (3.5–6.15) laminates when element miniaturization or sub-wavelength spacing is needed—for instance, to suppress grating lobes in a scanned-beam design. The RO3003G2 laminate is a strong middle-ground choice for Ku- and Ka-band reflectarrays, balancing low loss with moderate element size. Where miniaturization is paramount, the RO4360G2 high-Dk laminate can shrink patch dimensions by roughly 35 % compared to a Dk 3.0 substrate.
Substrate thickness affects both bandwidth and phase sensitivity. Thicker substrates (20–60 mil / 0.5–1.5 mm) give wider bandwidth but shallower phase curves, requiring finer dimensional control. A practical guideline: keep the substrate between 0.02 λ and 0.06 λ at the center frequency. For more on how radome layers interact with the element tuning, see the linked radome design guide—any cover placed over a reflectarray shifts element resonances and must be included in the unit-cell simulation.
Feed Configurations
Center-Fed vs. Offset-Fed
A center-fed reflectarray places the feed horn at the focal point directly above the array center. This gives symmetric illumination and the simplest phase distribution, but the feed and its support structure block part of the aperture—typically 1–3 % of the area, causing 0.2–0.5 dB of blockage loss and increased sidelobes.
An offset-fed arrangement tilts the feed to one side, eliminating blockage entirely. The trade-off is an asymmetric illumination taper that must be compensated in the element phasing, and a slightly more complex mechanical design. Offset feeding is standard for apertures larger than about 15 λ, where blockage loss becomes significant.
Feed Horn Selection
The feed should have a –10 dB beamwidth that matches the subtended angle of the array from the focal point. A corrugated conical horn or a wideband tapered-slot feed are common choices. For dual-polarization or circular polarization, a septum-polarized horn is typical at Ku/Ka-band.
Bandwidth Limitations and Broadband Techniques
Bandwidth is the chief weakness of reflectarrays. Two mechanisms limit it:
- Element bandwidth: Each patch has a finite resonant bandwidth over which its reflection phase is stable. A single-layer patch on a thin substrate may hold phase within ±20° over only 3–5 % bandwidth.
- Spatial phase delay: The path-length compensation designed at the center frequency becomes incorrect at other frequencies because the physical path lengths do not scale with wavelength. This effect worsens with aperture size; a rule of thumb is that the spatial-delay bandwidth is roughly λ / (2 · ΔL_max), where ΔL_max is the maximum differential path length across the aperture.
Broadband techniques include multi-layer stacked elements (12–15 % bandwidth), true-time-delay elements using coupled lines or aperture-coupled patches (up to 20 %), and sub-aperture partitioning. For applications demanding octave or multi-octave bandwidth, a transmitarray lens approach may be more practical, since transmitarrays can implement true-time-delay more naturally through multi-layer phase-shifting cells.
PCB Fabrication Considerations

Dimensional Tolerances
Reflectarray performance is directly tied to how accurately each element is etched. A patch length error of ±25 µm (±1 mil) on a Ku-band element shifts the reflection phase by roughly 5–15°, depending on the substrate and element type. For arrays targeting sidelobe levels below –25 dB, specify etch tolerance of ±15 µm (±0.6 mil) or better—achievable with modern photolithographic processes on copper-clad laminates.
Panel Size and Tiling
Large reflectarrays (> 500 mm per side) often exceed the maximum panel size of standard RF laminates (typically 457 × 610 mm or 610 × 915 mm for Rogers PTFE-based materials). The solution is tiling: fabricating sub-panels and assembling them on a flat backing structure. The gap between tiles must be kept under 0.05 λ to avoid grating-lobe artifacts. Alignment dowel pins and precision CNC-routed edges are standard practice. The transition from prototype to volume production requires careful attention to tile-to-tile phase continuity and bonding flatness.
Surface Finish
ENIG (electroless nickel / immersion gold) is the most common finish for reflectarray patches. The thin nickel layer (3–5 µm) adds a small amount of loss at Ka-band and above; immersion silver or bare copper with OSP can reduce this, at the cost of shelf life and solderability if active components are present.
Applications of Reflectarray Antennas
- Ku/Ka-band satcom terminals: Flat-panel reflectarrays are a leading architecture for satellite antenna PCB designs targeting LEO and GEO links, where a 30+ dBi gain panel can replace a 45 cm parabolic dish at a fraction of the depth.
- Automotive radar: At 77 GHz, a small reflectarray (50 mm × 50 mm) can achieve 25+ dBi for long-range imaging radar with a simple feed, avoiding the complexity of a fully active phased array.
- 5G mm-wave backhaul: Fixed-beam reflectarrays at 28 GHz or 39 GHz provide cost-effective high-gain links between base stations.
- Radio astronomy and deep-space: Deployable reflectarrays printed on flexible polyimide substrates have been demonstrated for CubeSat missions, folding into a 3U volume and deploying to a 0.5 m aperture.
Reflectarray vs. Phased Array vs. Parabolic Dish
| Parameter | Reflectarray | Active Phased Array | Parabolic Dish |
|---|---|---|---|
| Profile | Flat + feed standoff | Flat (conformal possible) | Deep curved surface |
| Beam steering | Fixed or mechanically tilted; reconfigurable with PIN/varactor elements | Electronic, fast | Mechanical gimbal |
| Bandwidth (1-dB gain) | 5–15 % | 20–40 %+ | Very wide (feed-limited) |
| Cost per aperture area | Low (passive PCB) | High (T/R modules) | Low (machined metal) |
| Weight | Low | Medium | High |
| DC power | Zero (passive) | High | Zero |
Choose a reflectarray when you need high gain, low profile, low cost and can tolerate moderate bandwidth and fixed (or slowly steered) beams. Choose an active phased array when fast electronic scanning and wide instantaneous bandwidth are required and the budget supports T/R modules. Choose a dish when volume and weight are unconstrained and wideband performance matters most.
Design Workflow Summary
- Specify requirements: frequency, gain, bandwidth, polarization, sidelobe level, mechanical envelope.
- Select substrate: Dk, Df, thickness, panel size, cost.
- Design the unit cell: full-wave simulation (Floquet boundary conditions in HFSS, CST or Ansys HFSS) sweeping the element dimension to extract the reflection phase and magnitude curves.
- Compute the phase map: from the feed position, beam direction and element grid, calculate the required phase at every cell.
- Map phase to geometry: assign each cell the dimension (or rotation) closest to its target phase using the unit-cell library.
- Full-array simulation: verify gain, pattern, sidelobes and cross-polarization. Iterate feed position and element grid if needed.
- Generate Gerber files: each element becomes a unique pad in the layout. Panelize for fabrication, specify etch tolerance, and include fiducials for tile alignment.
- Fabricate, assemble feed, and test: validate with antenna pattern measurement in a compact range or planar near-field scanner.
Frequently Asked Questions
What gain can a reflectarray antenna achieve?
Gains of 25–40 dBi are typical. A 30 cm × 30 cm aperture at Ku-band (12.5 GHz) delivers about 28–30 dBi with 50–60 % aperture efficiency. Larger apertures or higher frequencies push the number higher.
Can a reflectarray steer its beam electronically?
Yes, if the elements include tunable components—PIN diodes, varactor diodes or liquid-crystal cells. Each element’s reflection phase is then adjusted in real time. Beam-steering reflectarrays have been demonstrated at Ku-, Ka- and W-band, though the added component cost and bias network complexity partially offset the passive-array cost advantage.
Why is reflectarray bandwidth narrow compared to a dish?
Two effects combine: the resonant bandwidth of each patch element, and the frequency-dependent spatial phase error across the aperture. Multi-layer elements and true-time-delay techniques can extend the 1-dB gain bandwidth to 15–20 %, but the architecture is inherently more narrowband than a shaped reflector fed by a wideband horn.
What is the typical element spacing in a reflectarray?
Half a free-space wavelength (λ₀/2) is the standard starting point. At 12.5 GHz this is 12 mm; at 77 GHz it is about 1.95 mm. Sub-wavelength spacing (0.3–0.4 λ) is used when the beam must be scanned significantly off broadside or when grating-lobe suppression is critical.
Is FR-4 suitable for a reflectarray?
Only at low frequencies (below roughly 3 GHz) and for proof-of-concept work. Above 5 GHz the Dk variation (±0.3 typical) and high Df (0.020+) of standard FR-4 cause unacceptable phase errors and element losses. Production reflectarrays almost always use controlled-Dk RF laminates from Rogers RO4000 or RT/duroid families or equivalent grades from Isola’s Astra MT77 line.
Ready for a quotation?
Send your requirements for a material, stackup and manufacturing review. Price and schedule are confirmed for your project.
