In conventional wireless system design, a foundational assumption has been that adding more antennas to a base station yields a proportional increase in performance. Early theoretical frameworks for Massive MIMO suggested that the total captured energy or signal-to-noise ratio (SNR) scales linearly ($N$) with the number of antennas, while models for Intelligent Reflecting Surfaces (IRS) predicted an even more dramatic quadratic ($N^2$) power scaling law.
As Jakob Hoydis et al. noted, these early models often relied on the mathematical convenience of an infinite number of antennas distributed over an indefinitely expanding physical space. However, treating these scaling laws as unbounded creates a physical paradox: an infinitely growing SNR implies that a receiver could eventually capture more energy than the transmitter actually radiates, violating the fundamental law of conservation of energy.
As wireless research transitions toward 6G and extremely large-scale aperture architectures, a wave of recent literature has reconciled information theory with physical electromagnetics. The core insight is simple yet profound: as antenna arrays scale toward infinity, channel gain and SNR do not grow boundlessly; instead, they asymptotically converge to a finite upper bound.
The Geometry of the Near-Field
To understand why power gain saturates, we must look at how the physical relationship between a transmitter and receiver changes at scale. Traditional models rely on the "far-field" assumption, treating electromagnetic waves as plane waves that arrive parallel to one another across the entire array. This approximation holds when the array is small relative to its distance from the user.
However, when deploying Extremely Large-Scale MIMO (XL-MIMO) or continuous holographic surfaces, the physical size of the array expands dramatically. The user is suddenly operating within the "near-field" region, where wavefronts must be modeled accurately as spherical.
In Yong Zeng et al., the authors demonstrate that when spherical wavefronts are accounted for, the SNR hits a point of diminishing returns. This behavior is governed by the angular span (or solid angle) that the array subtends from the perspective of the user.
Imagine a point-source transmitter radiating a finite amount of power. The energy spreads outward spherically. An antenna array can only capture the fraction of energy that physically intersects its surface area. If you start with a small array and expand it, you initially capture more of the radiated energy. But as the array grows larger and larger, the peripheral elements become increasingly distant and tilted relative to the source. Eventually, expanding the array further adds almost no new energy capture because the array has already encompassed the maximum possible angular sector of the radiated signal.
Far-Field (Planar Waves) Near-Field (Spherical Waves)
| | | | \ | /
| | | | \ | /
============== ==============
[Linear Growth] [Angular Span Limit]
Mechanisms of Saturation: Aperture and Polarization Losses
The physical breakdown of unbounded scaling manifests through specific geometric and electromagnetic penalties as elements are added far from the array's center.
- Projected Aperture (Obliquity) Loss: As an array expands linearly, the elements at the far edges view the transmitter at an increasingly oblique angle. The effective physical area (the projected aperture) that these edge elements present to the incoming wave shrinks, reducing their ability to capture power.
- Increased Propagation Distance: Elements at the periphery of an extremely large array are physically farther from the user than the elements at the center. Because electromagnetic power decays with the square of the distance, these distant elements contribute progressively negligible amounts of power. Chao Feng et al. demonstrated this effect in the context of Extremely Large-Scale Intelligent Reflecting Surfaces (XL-IRS), proving that these geometric realities break the theoretical $N^2$ quadratic scaling law, forcing the received power to hit a hard physical ceiling.
- Polarization Mismatch: As detailed in a tutorial review by Zhaolin Wang et al., accurate near-field modeling must also account for polarization losses. The polarization vector of an electromagnetic wave changes direction as it propagates spherically. As an antenna array becomes extremely large, the polarization alignment between the transmitter and the peripheral antenna elements degrades, further accelerating the saturation of received power.
Universality Across Architectures
This asymptotic convergence is a universal physical constraint that applies regardless of how the large aperture is constructed.
Modular and Discrete Arrays
Whether an array is built as a single massive panel or split into distributed components, energy conservation remains absolute. Xinrui Li et al. analyzed modular XL-MIMO systems deployed across non-contiguous surfaces (like building facades). They derived closed-form expressions confirming that the SNR approaches a finite upper bound dictated strictly by the total energy physically radiated toward the total footprint of the modules.
Continuous and Holographic Surfaces
One might wonder if eliminating discrete elements in favor of a continuous, software-defined "holographic" surface could bypass these limits. Boqun Zhao et al. investigated this in Continuous-Aperture Array (CAPA) systems designed for Integrated Sensing and Communications (ISAC). They found that while communication and radar sensing performance scales with larger surface areas, the performance rates still converge to finite upper bounds. A continuous surface cannot capture more photons than pass through its spatial footprint.
The Role of the Reactive Near-Field
Finally, researchers have questioned whether non-propagating fields close to the antenna surface—the reactive near-field, home to evanescent waves—could inject extra gain into the system. Chongjun Ouyang et al. rigorously analyzed this region and proved that while the reactive region introduces complex local electromagnetic interactions, its practical impact on channel gain is negligible. Even when antenna density or aperture size approaches infinity, the channel gain still converges to the exact same finite constant limit.
Conclusion
The realization that power gain saturates changes how engineers must approach 6G network planning. Packing infinite antennas into a space or building infinitely large panels yields diminishing returns. Instead of chasing raw aperture size, the focus of next-generation design must shift toward optimizing the placement of arrays to maximize their angular coverage relative to users, ensuring that the physical footprint of the array is utilized efficiently within its asymptotic limits.
Key Foundational Work
How Does Performance Scale with Antenna Number for Extremely Large-Scale MIMO? (2020) — contributed 2 concept elements to this direction.
Massive MIMO: How many antennas do we need? (2011) — contributed 1 concept elements to this direction.