Static-Passive Electromagnetic Skin (SP-EMS)

Updated 8 July 2026

SP-EMS is a fully passive, static metasurface characterized by fixed meta-atom phase profiles that enable engineered beam shaping without runtime control.
It utilizes advanced electromagnetic synthesis methods, such as GSTC models and inverse-source techniques, to achieve functionalities like retro-reflection, focusing, and dual-polarization responses.
SP-EMS offers a cost-effective, low-complexity solution for enhancing coverage, localization, and ISAC performance in applications ranging from vehicle roofs to building facades and indoor walls.

Static-Passive Electromagnetic Skin (SP-EMS) denotes a fully passive, non-reconfigurable metasurface whose electromagnetic response is fixed at design or fabrication time and realized by the spatial arrangement of meta-atoms rather than by run-time control. Within the broader electromagnetic-skin taxonomy, SP-EMS is the static counterpart of RIS-like platforms: it manipulates reflection, retro-reflection, focusing, beam shaping, coverage, or refraction without embedded active elements, bias lines, or control circuitry, and it has been studied on vehicle roofs, building facades, indoor walls, and glass panels as a low-complexity component of smart electromagnetic environments, smart radio environments, and ISAC systems (Tagliaferri et al., 2023, Maleki et al., 2 Nov 2025, Oliveri et al., 2022, Oliveri et al., 2023).

1. Conceptual definition and taxonomy

An SP-EMS is described in the literature as a fully passive, static metasurface or smart skin whose local phase profile is predetermined and immutable during operation. “Static” means that the phase profile of each meta-atom is fixed and cannot be adapted to changing radar, base-station, or user positions. “Passive” means that the surface contains no active amplification, no phase shifters, no embedded RF or baseband electronics, and no external bias or control network; operation is driven only by the illuminating wave (Tagliaferri et al., 2023). In urban channel-charting work, the same concept appears under the labels “fully-passive EMS,” “static EMS,” and “pre-configured EMS,” with runtime-free operation and cost on the order of a few dollars per square meter (Maleki et al., 2 Nov 2025).

This definition places SP-EMS between conventional metallic reflectors and reconfigurable intelligent surfaces. Compared with flat conductive plates, SP-EMS can impose engineered phase gradients, focusing laws, or retro-reflective responses; compared with RIS, they relinquish run-time adaptability in exchange for substantially lower hardware complexity and cost. Vehicle-localization studies explicitly characterize SP-EMS as “two/three orders of magnitude less” expensive than RIS and much simpler to integrate on vehicles (Tagliaferri et al., 2023). A related adjacent class is the one-time programmable EMS: it is fabricated as a generic passive platform, configured once through expendable components such as fuses, and then operates as a passive-static, zero-maintenance skin, i.e., operationally as an SP-EMS after programming (Oliveri et al., 19 Jul 2025).

A common misconception is that “static” implies “single-function.” The recent multi-polarization literature shows that a single static-passive panel can simultaneously realize two independent reflection functionalities at the same frequency by exploiting TE/TM polarization diversity rather than electronic reconfiguration (Oliveri et al., 14 Aug 2025). Staticity therefore constrains temporal adaptability, not necessarily functional richness.

2. Electromagnetic principles and synthesis methodologies

The electromagnetic design of SP-EMS is usually expressed through generalized Snell phase gradients, GSTC sheet models, or inverse-source formulations. In vehicle retro-reflection, for example, a static module designed for incidence direction $\overline{\boldsymbol{\xi}_p}=[\overline{\theta}_p,\overline{\phi}_p]^T$ uses the fixed phase law

$\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$

which is the retro-reflection phase profile of a RIS-like aperture, but frozen at fabrication time (Tagliaferri et al., 2023).

At a more general level, GSTC-based formulations model the skin as a zero-thickness sheet whose electric and magnetic polarization densities generate equivalent surface currents and, in turn, the scattered field. A representative formulation writes the far field as

$\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$

with the currents linked to electric and magnetic surface polarizations through GSTC relations (Oliveri et al., 2021). This formalism underlies later generalized near-/far-field analysis, where Fresnel-region terms supplement standard array-phase factors to support both anomalous reflection and finite-distance focusing within a unified model (Oliveri et al., 2022).

Synthesis methodologies are correspondingly multiscale. Holographic and system-by-design approaches first specify a target field, power mask, beam, or focal footprint, then derive ideal current distributions, and finally map those currents to realizable local descriptors such as patch sizes or other meta-atom parameters (Oliveri et al., 2021). Aperiodic finite-size effects are a persistent issue for large static skins; AI-enhanced workflows therefore replace naive local-periodicity assumptions with digital twins trained on finite aperiodic sub-arrays, so that non-uniform mutual coupling and edge effects enter the surrogate mapping from unit-cell geometry to effective susceptibility tensors (Oliveri et al., 2021). For inexpensive single-layer implementations, inverse-source methods exploit the non-uniqueness of the radiation operator by decomposing the ideal current into pre-image and null-space parts,

$\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$

and then alternately optimizing the physical layout and the null-space coefficients to better match realizable currents on lossy substrates (Oliveri et al., 2024).

3. Architectures, materials, and representative embodiments

SP-EMS implementations span several hardware archetypes, but they share a recurring pattern: subwavelength or near-subwavelength periodicity, passive patterned metallization, and a geometry-specific reflection or transmission law computed offline and frozen in the fabricated panel.

Setting	Implementation	Representative attributes
Vehicle roof retro-reflector	Modular planar metasurface on Rogers3003 with square printed patches	Lattice $\approx 0.3\lambda_0$ at $\sim 78.5$ GHz; multiple fixed-incidence modules (Tagliaferri et al., 2023)
Building-facade coverage skin	Tile-based reflective coating on an admissible facade surface	Binary layout of installed/absent tiles optimized over the wall (Rocca et al., 2021)
Indoor Wi-Fi wall skin	Dual-layer, metal-backed patch panel on FR-4	$20\times20$ unit cells, half-wavelength spacing, side $0.55$ m at $5.64$ GHz (Benoni et al., 2023)
Transparent window skin	Meshed copper rings on standard insulating glass	26 GHz, optical transmittance $>80\%$ , non-Snell transmission (Oliveri et al., 2023)
Multi-polarization skin	Single-layer rectangular patches on Rogers RO3003	Independent TE/TM functions at 28 GHz with $\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 0 dB and phase coverage $\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 1 (Oliveri et al., 14 Aug 2025)

On vehicles, the surface is mounted on the roof because that region is most visible to radars located above the road and provides a relatively planar support. The proposed omnidirectional solution is not a continuously steerable surface but a set of $\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 2 static modules, each optimized for a different incidence direction, arranged so that only one module’s main lobe is aligned with the radar in a realistic geometry (Tagliaferri et al., 2023). On building facades, the architecture can instead be modular at the tile level: an admissible wall area is partitioned into square tiles, each either present or absent, and each installed tile is statically designed to redirect energy to a prescribed region of interest (Rocca et al., 2021).

Indoor SP-EMS demonstrations adopt monolithic PCB-like panels. In the large-scale Wi-Fi experiment, the hallway skin is a $\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 3 m square aperture built from a $\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 4 array of dual-layer, metal-backed square patches on FR-4, with patch sizes chosen to realize a fixed anomalous reflection law toward a low-coverage region (Benoni et al., 2023). Transparent outdoor-to-indoor skins take a different route: meshed copper rings are attached through optical clear adhesive to standard insulating glass, so that the window itself becomes the effective support of a transmissive static EMS with visible-light transmittance above $\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 5 and engineered non-Snell refraction (Oliveri et al., 2023).

A separate branch targets multifunctionality rather than minimal geometry. Multi-polarization SP-EMS uses a single rectangular-patch meta-atom on a square lattice at 28 GHz; by exploiting polarization-dependent reflection coefficients, the same fixed aperture supports one TE reflection law and another, independent TM reflection law (Oliveri et al., 14 Aug 2025).

4. Functional roles in wireless systems, sensing, and localization

The canonical system role of SP-EMS is coverage shaping in shadowed or weak-service regions. Facade-mounted reflective skins are optimized so that one or more static beams illuminate an area of interest where direct base-station power is below threshold. In the modular facade formulation, the optimization problem jointly minimizes coverage deficit and layout complexity, thereby selecting a minimum-cardinality subset of tiles that restores or improves service in urban blind spots (Rocca et al., 2021). At city scale, the same logic is embedded in planning frameworks that treat candidate walls as binary decisions and use digital twins to decide where low-cost passive static reflective skins should be installed for QoS recovery (Benoni et al., 2021).

Indoor SP-EMS uses the same principle in a more controlled geometry. A wall-mounted skin intercepts part of the field radiated by an existing access point and re-radiates it in a pre-defined anomalous direction toward a target region of interest in a hallway, leaving AP positions, transmit power, and cabling unchanged (Benoni et al., 2023). Transparent window skins extend that idea to outdoor-to-indoor mmWave links: instead of reflecting, they transmit through existing glass panels while imposing a phase gradient that redirects the incident field into the building along a non-Snell direction, thereby reducing the effective penalty of standard glass penetration (Oliveri et al., 2023).

A second major role is target marking and sensing simplification. In vehicular localization, a bare vehicle illuminated by a wideband automotive radar behaves as an extended target with multiple aspect-dependent scattering centers, double bounces, layover, and shadowing. A roof-mounted SP-EMS acts as a compact high-reflectivity planar retro-reflector placed at the desired localization point, so that the radar tracks a single stable rooftop marker rather than arbitrary bright points on wheels, bumpers, or ground interactions. This “pointization” suppresses perspective-induced ranging bias and simplifies tracking and data association (Tagliaferri et al., 2023).

A third role is propagation shaping for data-driven localization. In channel charting, the relevant objective is not maximum received power per se but a balance between SNR and channel dissimilarity. Fully passive, static EMS can add scenario-tailored multipath that preserves or enhances location fingerprints, whereas RIS-like gain-maximizing strategies may collapse dissimilarity between nearby and far users and thereby degrade chart quality. This makes SP-EMS particularly suitable when the radio environment itself is used as a geometric sensing substrate (Maleki et al., 2 Nov 2025).

Finally, SP-EMS can exceed the capabilities of flat passive conductive screens in NLOS specular links. Because a patterned skin can impose a focusing phase distribution rather than merely mirror the incident wave, it can behave as a passive focusing aperture and improve the link budget beyond the asymptotic image-theory limit associated with an infinite flat reflector (Oliveri et al., 2022).

5. Performance regimes and representative results

Several performance regimes recur across the literature. The first is the distinction between far-field beam steering and radiative-near-field focusing. Generalized analysis shows that the ideal current phase for a target point at finite $\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 6 contains both a linear steering term and a quadratic Fresnel term; the former dominates in the far field, while the latter becomes essential for near-field focusing (Oliveri et al., 2022). In the vehicle case, a related wideband effect appears as beam squint and spatial filtering: the EMS array factor varies over GHz-wide bandwidths, so the skin can act as a spatially wideband filter rather than a frequency-flat reflector (Tagliaferri et al., 2023).

The NLOS-link literature provides a compact ideal benchmark. For an ideal EMS screen of side $\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 7 in a specular two-hop geometry, the upper-bound total path attenuation is

$\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 8

which implies quartic growth with side length and leads to the threshold condition

$\Phi_{p,nm}(\overline{\boldsymbol{\xi}_p}) = \frac{4\pi f_0}{c}\, d \left( n\sin\overline{\phi}_p\cos\overline{\theta}_p + m\sin\overline{\phi}_p\sin\overline{\theta}_p \right),$ 9

for outperforming the asymptotic PCS limit (Oliveri et al., 2022). In a representative 27 GHz example with $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 0 m, $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 1, and $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 2 m, the implemented SP-EMS achieved about $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 3 dB improvement over a same-size PCS, about $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 4 dB over the asymptotic PCS limit, and remained about $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 5 dB below the ideal EMS upper bound (Oliveri et al., 2022).

For localization, the main figures of merit are RCS, CRB, and PEB. Roof-mounted EMS make vehicles behave like point targets in slant range and can improve PEB by an order of magnitude relative to a bare vehicle, depending on module size and bandwidth (Tagliaferri et al., 2023). The same paper reports HFSS validation of modules of size $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 6 cm $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 7 and $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 8 cm $\mathbf{E}^{FF}(\mathbf{r}) \approx \frac{j k_0}{4\pi} \frac{\exp(-j k_0 |\mathbf{r}|)}{|\mathbf{r}|} \int_{\Omega} \left\{ \hat{\mathbf{r}} \times \left[ \eta_0 \hat{\mathbf{r}} \times \mathbf{J}^e(\tilde{\mathbf{r}}) + \mathbf{J}^m(\tilde{\mathbf{r}}) \right] \exp\big(j k_0 \hat{\mathbf{r}}\cdot \tilde{\mathbf{r}}\big) \right\} \, d\tilde{\mathbf{r}},$ 9 over 76–81 GHz, with the smaller module retaining the design elevation within the $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 0 dB beam over the full band, which is interpreted as limited spectral filtering and is favorable for wideband localization (Tagliaferri et al., 2023).

For channel charting, the headline result is a reduction of the $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 1th-percentile localization error from $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 2 m to $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 3 m in a 3D ray-traced 30 GHz city, consistently for semi-supervised t-SNE and autoencoder pipelines, together with a reduction of severe trajectory dropouts by more than $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 4 under $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 5 supervision (Maleki et al., 2 Nov 2025). The underlying design principle is that optimal static EMS configurations are neither pure SNR maximizers nor random-phase maximizers: they occupy an intermediate region in the SNR–dissimilarity trade-off that preserves chart geometry (Maleki et al., 2 Nov 2025).

Transparent OTO-EMS demonstrates a different metric set. At broadside transmission through glass, the designed skin yields about $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 6 dB improvement over the glass-only case at $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 7 and is only about $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 8 dB below the hollow-aperture reference; for a $\widetilde{\mathbf{J}}(\mathbf{r}) = \mathbf{J}_{PI}(\mathbf{r}) + \mathbf{J}_{NS}(\mathbf{r}),$ 9 scan, the scan loss is about $\approx 0.3\lambda_0$ 0 dB relative to the broadside beam of the same EMS (Oliveri et al., 2023). Multi-polarization SP-EMS, by contrast, is assessed mainly through unit-cell efficiency and beam fidelity: the meta-atom maintains $\approx 0.3\lambda_0$ 1 dB with phase coverage $\approx 0.3\lambda_0$ 2, and synthesis time scales from about $\approx 0.3\lambda_0$ 3 s for a $\approx 0.3\lambda_0$ 4 panel to about $\approx 0.3\lambda_0$ 5 s for a $\approx 0.3\lambda_0$ 6 panel in non-optimized single-core MATLAB (Oliveri et al., 14 Aug 2025).

Indoor experimental evidence is particularly explicit. In hallway “A,” a $\approx 0.3\lambda_0$ 7 m SP-EMS increased measured average received power by $\approx 0.3\lambda_0$ 8 dB, achieved a measured maximum local gain of $\approx 0.3\lambda_0$ 9 dB, reduced the low-coverage region from $\sim 78.5$ 0 m $\sim 78.5$ 1 to $\sim 78.5$ 2 m $\sim 78.5$ 3, and reduced the fraction of points below $\sim 78.5$ 4 dBm from $\sim 78.5$ 5 to $\sim 78.5$ 6 (Benoni et al., 2023).

6. Deployment economics, limitations, and research directions

The economic rationale for SP-EMS is as prominent as the electromagnetic one. Because static skins eliminate tunable elements, control buses, drivers, and run-time optimization, they are repeatedly presented as a mass-deployment alternative to RIS. Vehicle work places the cost gap at two to three orders of magnitude in favor of SP-EMS over RIS (Tagliaferri et al., 2023), while urban channel-charting work reports a cost on the order of a few dollars per square meter for fully-passive static EMS (Maleki et al., 2 Nov 2025). In the indoor Wi-Fi testbed, the hallway-A SP-EMS had acquisition and commissioning costs of \$\sim 78.5$75, respectively, zero operational cost per year, and total cost \$\sim 78.5$81350 over five years; the resulting saving was about $\sim 78.5$9, or about \$20\times20$0^2$ over the hallway area (Benoni et al., 2023).

The main limitation is the same property that defines the class: absence of run-time adaptability. A static skin cannot correct misalignment, react to traffic changes, or retune for a new incidence geometry. For vehicle roofs this creates an angular-coverage problem: to achieve full azimuth coverage with elevation span, the cited estimate is about 216 modules of size $20\times20$ 1 cm $20\times20$ 2, occupying a roughly 60 cm side area on the roof (Tagliaferri et al., 2023). For channel charting, performance remains strongly scenario-dependent, with placement and codebook choice materially affecting the outcome (Maleki et al., 2 Nov 2025). Transparent window skins inherit phase-coverage and insertion-loss limitations from the two-layer optically transparent implementation, so performance degrades away from the design frequency and at extreme scan angles (Oliveri et al., 2023).

A second limitation is that low complexity often trades against controllability. The very inexpensive paper-and-conductive-ink SP-EMS studied in inverse-source work exhibits poor phase linearity and large reflection-magnitude variation with patch size; the null-space optimization improves performance, but the platform remains narrower-band and less efficient than higher-grade PCB implementations (Oliveri et al., 2024). Likewise, static facade and city-scale planning methods depend on accurate digital twins, ray-tracing data, or geometry-aware candidate-wall selection, so model mismatch can directly propagate into suboptimal installations (Benoni et al., 2021, Rocca et al., 2021).

Current research directions follow two trajectories. One keeps the passive-static operating principle but broadens functionality: multi-polarization static skins, transparent window skins, modular facade tilings, and one-time programmable passive skins that preserve zero-maintenance operation after installation-time configuration (Oliveri et al., 14 Aug 2025, Oliveri et al., 19 Jul 2025). The other trajectory seeks tighter integration with broader network design, including joint BS/EMS planning, multi-panel optimization, hybrid static/dynamic smart environments, and standardization questions concerning interoperability, regulation, and large-scale deployment in 6G smart radio environments (Benoni et al., 2021, Rocca et al., 2021).

Taken together, the literature portrays SP-EMS not as a reduced RIS, but as a distinct design point: a precomputed electromagnetic artifact whose value comes from low cost, structural integrability, and physically encoded field transformation. Where geometry is stable and infrastructure-scale deployment matters, that trade-off is not incidental; it is the central engineering premise.