Visible Light Communication (VLC) systems have emerged as a promising wireless technology that leverages light-emitting diodes (LEDs) to transmit data by modulating light intensity at speeds imperceptible to the human eye. As the demand for high-speed data transfer surges alongside the proliferation of Internet of Things (IoT) devices, smart lighting, and autonomous vehicles, understanding the fundamental capacity limits of VLC systems becomes critical. This article provides an in-depth exploration of the factors that constrain VLC capacity, the current state of achievable data rates, and the innovative strategies researchers are developing to push beyond these boundaries. By examining the interplay of hardware constraints, channel characteristics, and modulation techniques, we gain a comprehensive view of how VLC can complement or even replace conventional radio frequency (RF) communication in specific environments.

Fundamentals of VLC Technology

VLC operates by encoding data onto the rapid modulation of light intensity from LEDs. A photodetector at the receiver converts these optical intensity variations back into an electrical signal, which is then demodulated to recover the transmitted information. The core advantage of VLC lies in its use of the vast, unregulated optical spectrum (400–800 THz), which offers immunity to RF interference, inherent security due to signal confinement by opaque barriers, and the potential for high spatial reuse. Early VLC systems achieved modest data rates (tens of Mbps), but modern implementations using sophisticated modulation and multiple-input multiple-output (MIMO) techniques have demonstrated gigabit-per-second speeds in laboratory settings.

The channel capacity of a VLC system is ultimately bounded by Shannon's information theory, adapted for the unique characteristics of the optical wireless channel. Unlike RF channels that are often described by additive white Gaussian noise (AWGN), VLC channels are subject to shot noise from ambient light, thermal noise in the receiver, and the bandlimited nature of both the LED and photodiode. Moreover, the channel exhibits a low-pass frequency response, with the 3 dB bandwidth of typical phosphor-converted white LEDs often limited to a few megahertz. Overcoming this bandwidth bottleneck is a central theme in capacity research.

Key Factors Governing VLC Capacity

Bandwidth of LED Sources

The modulation bandwidth of an LED is determined by the minority carrier lifetime in the semiconductor material. Standard high-brightness white LEDs, which use a blue LED chip exciting a yellow phosphor, have a bandwidth of only 2–20 MHz due to the slow phosphor response. Micro-LEDs (μLEDs) and resonant-cavity LEDs (RC-LEDs) can achieve bandwidths exceeding 100 MHz, while laser diodes (LDs) offer even higher bandwidths (multiple gigahertz) but at higher cost and complexity. The choice of light source directly sets an upper bound on the achievable data rate without advanced equalization or spectral slicing.

Modulation and Coding Techniques

The capacity limit is strongly influenced by the modulation format. Simple on-off keying (OOK) is easy to implement but inefficient. Advanced schemes such as orthogonal frequency-division multiplexing (OFDM), pulse-amplitude modulation (PAM), and carrierless amplitude and phase (CAP) modulation significantly improve spectral efficiency. OFDM, in particular, has been widely adopted because it can mitigate intersymbol interference (ISI) caused by the channel's low-pass response and can be adapted to avoid DC wander. However, OFDM signals have a high peak-to-average power ratio (PAPR), which can drive LEDs into nonlinear regions, distorting the signal and reducing effective capacity. Pre-distortion and adaptive bit loading can help, but they add complexity.

Signal-to-Noise Ratio (SNR)

The SNR in a VLC link is determined by the received optical power, the photo-detector's responsivity, and the noise sources. Shot noise from background light—especially in indoor environments with sunlight or other artificial lighting—can severely degrade SNR. Thermal noise in the receiver's front-end amplifier also imposes a floor. The capacity scales logarithmically with SNR, so improving the link budget through better optics, higher-power LEDs, or reduced ambient interference directly increases capacity. In practice, the SNR in VLC is often lower than in RF links due to limited transmit power for eye safety and the path loss from diffuse reflections.

Channel Characteristics and Line-of-Sight

VLC channels can be line-of-sight (LOS), non-line-of-sight (NLOS) via reflections, or a mix. LOS links provide the highest power efficiency and bandwidth, but require unobstructed paths. NLOS links rely on diffuse reflections from walls and ceilings, which extend coverage but introduce severe multipath dispersion, limiting the data rate to tens of Mbps. The channel's delay spread in NLOS scenarios can be tens of nanoseconds, necessitating complex equalization or OFDM with a cyclic prefix. The trade-off between coverage and capacity is a fundamental design consideration.

Current State-of-the-Art and Achievable Rates

Research has demonstrated impressive progress. In 2020, a team achieved 15 Gbps using a single GaN micro-LED with OFDM, while MIMO systems with arrays of micro-LEDs have reached beyond 100 Gbps in controlled settings. Commercial Li-Fi systems (a subset of VLC) now offer up to 1 Gbps per access point. However, these results often require highly optimized components, careful alignment, and minimal interference. Practical deployments face additional hurdles: dimming requirements, flicker constraints, and the need to support multiple users. The theoretical capacity of a VLC channel under practical constraints is still an active research area, with many open questions regarding the limits imposed by LED nonlinearity and the optical multipath channel.

Strategies to Push Capacity Boundaries

Multiple-Input Multiple-Output (MIMO)

MIMO VLC uses arrays of LEDs at the transmitter and multiple photodetectors at the receiver to create spatial channels. Unlike RF MIMO, where independent fading paths are easily exploited, VLC MIMO channels are often highly correlated because all LEDs emit light into the same optical path. Techniques such as imaging receivers (e.g., using a lens to create separate images of each LED) or angle-diversity receivers can de-correlate the channels. Spatial modulation, where different LEDs are activated based on the data bits, also offers a way to increase capacity without requiring many simultaneous RF chains.

Wavelength Division Multiplexing (WDM)

By using multiple LED colors (red, green, blue) or even narrowband laser sources, WDM can multiply the capacity by the number of independent wavelength channels. Each color channel is modulated separately and demultiplexed at the receiver using optical filters. Phosphor-converted white LEDs already contain a blue component and a broad yellow component, which can be separated to create two channels. Using discrete RGB LEDs, researchers have achieved multigigabit rates. The limitation is the availability of high-bandwidth LEDs at each wavelength and the complexity of optical filtering.

Advanced Signal Processing and Equalization

To overcome the LED's low-pass bandwidth, equalization techniques such as decision-feedback equalization (DFE) or Tomlinson-Harashima precoding can compensate for the channel's frequency response. OFDM with adaptive bit loading assigns higher-order modulation to subcarriers with better SNR, maximizing throughput. Nonlinear equalizers, including Volterra series-based models, address LED nonlinearity, which becomes a major capacity bottleneck at high optical powers. Machine learning algorithms are also being explored to optimize constellation shapes and pre-distortion in real time.

Optical Fiber-Like Approaches

Integrating VLC with optical fiber infrastructure allows remote modulation and distribution of light using fiber-fed remote heads. This approach can achieve very high bandwidth by using external modulators (e.g., Mach-Zehnder) that bypass the LED's bandwidth limit, directly modulating a laser source. The light is then coupled into free space via a diffuser or lens. This method has demonstrated capacities exceeding 100 Gbps over short ranges, but it sacrifices the simplicity and low cost of conventional LED-based VLC.

Integration with RF and Other Technologies

Hybrid systems that combine VLC with RF (such as Wi-Fi or millimeter-wave) can exploit the strengths of each: VLC provides high capacity in short-range, line-of-sight situations, while RF handles mobility and non-line-of-sight coverage. Such heterogeneous networks require intelligent handover mechanisms and resource allocation. Visible-light-based positioning (VLIP) can also be integrated for indoor navigation without additional hardware.

Advanced Materials and Device Design

New semiconductor materials like gallium nitride (GaN) on silicon substrates are enabling micro-LED arrays with bandwidths over 1 GHz. Nanophotonic devices, such as photonic crystal LEDs or quantum dot LEDs, promise even higher efficiencies and modulation speeds. On the receiver side, avalanche photodiodes (APDs) and silicon photomultipliers offer higher sensitivity, improving the SNR-limited capacity.

Non-Orthogonal Multiple Access (NOMA)

NOMA allows multiple users to share the same time-frequency resource by superimposing their signals in the power domain. In VLC, this can increase the aggregate capacity in a multi-user scenario. However, the near-far problem and the need for successive interference cancellation add complexity. Research is ongoing to adapt NOMA to the peculiarities of the optical channel.

VLC's capacity limits are also being explored in challenging mediums like water, where RF fails. Blue-green wavelengths have low attenuation, allowing moderate data rates (hundreds of Mbps) over tens of meters. The capacity is limited by strong scattering and absorption, as well as turbulence. Similar strategies—MIMO, adaptive modulation, and equalization—are being applied to underwater optical wireless communication.

Conclusion

Visible Light Communication stands at a promising crossroads where fundamental capacity limits are being aggressively challenged by innovations in devices, signal processing, and system architecture. Understanding the interplay between LED bandwidth, nonlinearity, channel path loss, and noise is essential for designing practical systems that can deliver gigabit-per-second speeds in real-world environments. While laboratory results already surpass 100 Gbps, commercial adoption requires cost-effective solutions that balance capacity, range, and integration complexity. As research continues into MIMO, WDM, hybrid networks, and advanced materials, VLC will increasingly become a viable complement to RF wireless, especially in crowded spectrum environments or where security and interference avoidance are paramount. The ongoing exploration of VLC capacity limits not only advances the field of optical wireless communications but also paves the way for future high-speed, spectrum-efficient, and secure data networks.

External References: