Reflectarray-Based RIS Technology

• Reflectarray-based RIS is a programmable electromagnetic surface composed of tunable subwavelength elements that modulate phase and amplitude for customizing wireless channels.
• The design leverages varactor diodes, PIN/MEMS switches, and calibrated bias networks to achieve precise beamforming, low-power operation, and high scalability.
• Advanced architectures combine passive and active elements to enhance performance in beam steering, coverage extension, and MIMO channel improvements.

A reflectarray-based reconfigurable intelligent surface (RIS) is a programmable electromagnetic structure composed of a two-dimensional (or, recently, three-dimensional) array of tunable subwavelength elements, each capable of imposing a controlled phase—and, with advanced architectures, amplitude—modulation on incident electromagnetic waves. Reflectarray-based RISs are foundational to the practical realization of programmable wireless environments, providing fine-grained, low-profile, and energy-efficient means for beamforming, signal manipulation, and channel customization in wireless networks, radar, and sensing. The architecture fuses concepts from planar antenna arrays, metasurfaces, and electronic phase shifters, enabling low-power, near-passive control over propagation with scalability to hundreds or thousands of elements.

1. Fundamentals of Reflectarray-Based Reconfigurable Intelligent Surfaces

A reflectarray-based RIS is formed from an $M \times N$ lattice of electrically small unit cells, each integrated with a tunable circuit—most commonly varactor diodes, PIN diodes, or MEMS switches—connected to a patch or slot antenna on a dielectric substrate, with an underlying ground plane for full EM isolation (ElMossallamy et al., 2020, Huang et al., 2022, Gros et al., 2021). Each cell is digitally or analogically biased to achieve a programmable local reflection coefficient $\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$, where $A_{n,m}$ is the amplitude (typically near unity, barring unavoidable losses) and $\phi_{n,m}$ is the programmable phase, typically spanning $2\pi$ for varactor-based (continuous) or $b$-bit discrete precision for PIN/MEMS-based designs.

The physical operation relies on engineered resonance: by varying the bias of embedded diodes or MEMS loads, the cell’s effective electrical length or boundary condition is changed, thus shifting its reflection phase. A network of high-impedance, electromagnetically isolated bias lines delivers the tuning signals (ElMossallamy et al., 2020, Fara et al., 2021). The total aperture acts as a digitally controlled phased array, leveraging the phased reflectarray principle to shape, steer, or split the reflected wavefront.

Design of the bias network, control electronics, and calibration (including element non-ideality, substrate inhomogeneity, and mutual coupling compensation) is critical for system performance and is typically accomplished with a combination of lookup-table calibration, hierarchical control logic or codebooks, and over-the-air feedback (Sayanskiy et al., 2022, Ratajczak et al., 2023).

2. Electromagnetic Modeling and Path-Loss Behavior

The modeling of reflectarray-based RISs involves characterizing both the unit cell and full-aperture response. The local response is captured by the cell’s equivalent impedance, such as $Z_{\mathrm{load}}(V)$, and its reflection $\Gamma(\omega,V)$, arising from superposition of patch resonance and load reactance (ElMossallamy et al., 2020, Droulias et al., 2022). At the array level, the spatially dependent phase law assigned to the RIS elements dictates the anomalous reflection, i.e., direction, focus, or field shaping of the reflected beam.

End-to-end system modeling includes element patterns, beamforming array factors, and mutual coupling. In the far field, for an RIS of area $A$ and elements spaced by $d\leq\lambda/2$, the classic path-loss scaling law is

$\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$0

where $\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$1 and $\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$2 are the source-to-RIS and RIS-to-receiver ranges, and $\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$3 is the cell’s power efficiency (Ellingson, 2019, Huang et al., 2022). This scaling demonstrates the $\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$4 (aperture squared) power gain in the specular (beamformed) direction compared to a direct path, conditional on all cell phases precisely compensating path delays.

RIS-enabled links fundamentally exhibit cascaded channel behavior and require precise phase alignment (focusing vs. simple beamforming) for maximum gain, especially in the near-field or for large apertures (Ellingson, 2019, Droulias et al., 2022). Comparisons with specular reflection and continuous sheet models reveal that discrete reflectarray RISs can asymptotically achieve the theoretical (perfect-plate) enhancement provided design rules—such as half-wavelength element spacing, well-controlled element pattern, and calibrated phase—are observed (Ellingson, 2019, Droulias et al., 2022).

3. Phase Control Granularity: Continuous, Multi-Bit, and Binary Architectures

Reflectarray RISs are implemented with a range of phase-quantization schemes, directly influencing beamforming precision, sidelobe levels, and complexity.

• Continuous phase control: Achieved with varactor-tuned elements, offering nearly $\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$5 sweep with fine granularity ($\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$61° with high-resolution DACs). This approach eliminates quantization lobes and achieves near-theoretical gain but increases analog biasing complexity (Ratajczak et al., 2023, Fara et al., 2021).
• Multi-bit phase control: PIN diodes or MEMS switches can realize $\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$7-bit discrete phase shifts (commonly $\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$8–$\Gamma_{n,m} = A_{n,m} e^{j\phi_{n,m}}$9). Two-bit designs (0, 90, 180, 270°) provide most of the gain of continuous phase (typically within 1 dB), while reducing power and hardware complexity (Wu et al., 2024, ElMossallamy et al., 2020).
• Binary-phase RISs: Simplest configuration; only two phase states per element (0/π). Such schemes offer rapid, low-power actuation, and are shown to achieve $A_{n,m}$088% broadside aperture efficiency, with 1.2–1.8 dB loss compared to continuous phase, and higher sidelobe floors (Gros et al., 2021, Sayanskiy et al., 2022).

The optimal choice depends on application constraints: continuous or multi-bit is favored for demanding beam-shaping or minimizing quantization noise, while binary is effective when area, power, or reconfiguration speed are paramount.

4. System Architectures: Passive, Active, and Advanced Topologies

The overwhelming majority of reflectarray-based RIS designs to date are passive, relying on resonant elements and reconfigurable loads to steer or shape reflected energy with negligible power consumption—a key enabler for low-maintenance, scalable deployment in 6G systems.

Recent hybrid and active architectures, such as the amplifying and filtering RIS (AF-RIS) of Wu et al. (Wu et al., 2024), incorporate on-board power amplifiers and bandpass filters:

• AF-RIS elements perform frequency-selective signal amplification by routing x-polarized incident waves through shared Wilkinson-inspired filters/amplifiers, then splitting the energy across multiple phase-tuned re-radiation paths. The key innovations are 8× hardware reduction in amplifier count (sub-connected topology), >20 dB in-band gain, >20 dB out-of-band suppression, and digital 2-bit phase control for robust beam steering.
• Such architectures address the double-hop path loss and out-of-band interference limitations of passive RISs, enabling compact, spectrum-safe RIS-assisted relays and robust coverage extension, especially where dense spectral environments or link budgets preclude passive-only solutions (Wu et al., 2024).

Alternative architectures include 3D RIS cubes with multi-face coverage and binary amplitude gating for wide-angle, spatially multiplexed beam steering (Wang et al., 13 Feb 2026), and non-diagonal phase shift matrices enabling signal routing flexibility and maximized rank/channel gain (Li et al., 2022).

5. Beamforming, Channel Modeling, and Performance in Wireless Networks

Reflectarray-based RISs are analytically modeled as programmable mirrors with beam direction and shape dictated by imposed phase masks: $A_{n,m}$1 for focusing, or as a linear phase ramp for specular steering. The array factor determines mainlobe direction and sidelobe structure (Sayanskiy et al., 2022, Ratajczak et al., 2023, Razavizadeh et al., 2020).

In wireless networks, the impact of a RIS is realized via cascaded channels $A_{n,m}$2, $A_{n,m}$3, composed via the configuration of $A_{n,m}$4 (ElMossallamy et al., 2020, Huang et al., 2022). End-to-end channel hardening is observed as the number of elements grows, with Rician $A_{n,m}$5-factors 10–20 dB typical under well-aligned beamforming, approaching deterministic channel gain as $A_{n,m}$6 (Huang et al., 2022).

System-level simulation in multi-cell deployments demonstrates:

• Power and SINR gains scaling with $A_{n,m}$7 for coherent aperture, translating to 3–6 dB SINR improvement for large-format (e.g., 40$A_{n,m}$840-element) panels (Gu et al., 2022).
• Sidelobe management, coverage extension, and near/far-field focusing are intimately tied to phase law accuracy, element spacing (typically $A_{n,m}$9 to suppress grating lobes), and aperture area (Ratajczak et al., 2023, Gros et al., 2021, Ellingson, 2019).
• Joint optimization of RIS phase mask and base station beam steering yields SNR gains of 4–10 dB over random/no-RIS configurations in 3D deployment scenarios (Razavizadeh et al., 2020).
• In multi-user and MIMO settings, spatial multiplexing and channel rank enhancement are achievable by synthesizing independent multi-paths via deliberate RIS geometry and phase coding (Huang et al., 2022).

Reflectarray-based RISs have been experimentally validated as effective relays, access-point extenders, and ambient backscatter enhancers, with measured directivity gains of 25–30 dBi in mmWave (Gros et al., 2021), beam-scanning up to ±60° (Ataloglou et al., 8 Apr 2025), and communication-grade EVM improvements of 6–7 dB (Wang et al., 13 Feb 2026).

6. Energy Efficiency, Practical Implementation, and Hardware Trade-offs

Energy consumption of traditional, passive reflectarray RISs is orders of magnitude lower than that of active relays—limited largely to control electronics and static diode bias (nW–mW total for arrays of $\phi_{n,m}$01k cells) (Ratajczak et al., 2023, Sayanskiy et al., 2022). Advanced active RISs, by virtue of sub-connected or shared-amplifier architectures, maintain modest power draw (e.g., 0.84 W for a 4$\phi_{n,m}$18 AF-RIS) while achieving functional parity with passive RISs 10$\phi_{n,m}$2 their area (Wu et al., 2024).

Hardware trade-offs are governed by:

• Element design: Miniaturization (cell $\phi_{n,m}$3 $\phi_{n,m}$4) for evanescent-wave coupling and surface-wave beamforming (Ataloglou et al., 8 Apr 2025); element Q and resonance bandwidth; ohmic/dielectric losses.
• Phase resolution: Higher bits improve beamforming but increase circuit complexity and DC bias requirements; binary/1-bit offers simplicity but imposes higher quantization sidelobes (Sayanskiy et al., 2022, Gros et al., 2021).
• Bias network: Distributed, modular control (optical IR, microcontroller-per-block) versus centralized, depending on deployment scale and codebook reconfiguration requirements (Sayanskiy et al., 2022).
• Scalability: Hierarchical and distributed address/control schemes permit plug-and-play extensibility and reduce recabling overhead (Sayanskiy et al., 2022, Ratajczak et al., 2023).

Environmental robustness (to temperature, humidity, and mechanical deformation), mutual coupling mitigation, and over-the-air calibration are active engineering research topics, along with further reduction of control/feedback overhead in large arrays (ElMossallamy et al., 2020, Ratajczak et al., 2023, Droulias et al., 2022).

7. Advanced Applications and Future Directions

Reflectarray-based RISs are established enablers for diverse applications:

• Wireless relays and cell extenders: Compact AF-RIS designs that merge wideband, high-gain, and spectral-selectivity, with practical coverage extension without aperture scaling (Wu et al., 2024).
• 3D and volumetric coverage: Multiface, subarrayed cube architectures supporting full angular coverage and inter-surface transmission, enabling blind-spot-free EM control (Wang et al., 13 Feb 2026).
• Radar sensing: Electronically steered RISs as secondary coherent apertures for SNR and detection-probability boosting in target detection, with theoretical SNR improvements scaling as $\phi_{n,m}$5 (Buzzi et al., 2021).
• Ambient backscatter and IoT: RISs with continuous phase tuning achieving substantial BER reductions and energy-per-bit savings in batteryless sensor networks (Fara et al., 2021).
• Flexible channel engineering: Use of non-diagonal phase matrices for channel-rank improvement and spatial-multiplexed MIMO gains (Li et al., 2022).
• Surface-wave assisted beamforming: Architectures exploiting excited evanescent fields for amplitude and pattern shaping on subwavelength arrays, facilitating sector-beam optimization (Ataloglou et al., 8 Apr 2025).

Prospective research directions encompass co-design of RIS and network protocol stacks, real-time channel/information-theory adaptive phase control, wideband and broadband element development, and the integration of surface-based signal processing and edge intelligence (ElMossallamy et al., 2020, Wu et al., 2024). Experimental validation at mmWave and THz bands, large-area deployment, and cross-layer performance measurements remain central to the maturation of reflectarray-based RIS technology for 6G and beyond.