Significance Spatiotemporal-coding metasurfaces (STCMs) extend digital-coding metasurfaces by introducing periodic temporal modulation into programmable subwavelength arrays. This added degree of freedom allows an incident carrier to be redistributed among harmonics whose amplitudes, phases, polarizations, and radiation directions can be controlled through the coding sequence. The same aperture can therefore perform signal modulation, frequency conversion, beam formation, and channel allocation. These capabilities are relevant to wireless systems that require compact radio-frequency (RF) front ends, agile spectrum use, multiuser transmission, or joint communication and sensing. A review focused on communication is needed because the field spans electromagnetic theory, programmable hardware, modulation methods, and system-level demonstrations, while the relationship between harmonic control and communication functions is often described separately.

Progress This review first traces the transition from static digital-coding metasurfaces to time-domain and spatiotemporal coding. Static programmable metasurfaces map discrete unit states to amplitude, phase, polarization, or far-field patterns. They have enabled direct digital modulation (DDM), dual-band amplitude coding, and RF-chain-free transmitters. Representative systems include an eight-phase-shift-keying (8PSK) transmitter operating at 4.25 GHz with a reported data rate of 6.144 Mbps, as well as binary- and quadrature-phase-shift-keying schemes implemented by changing the bias-dependent reflection coefficient. These studies established that a programmable aperture can participate directly in modulation and radiation, but its ability to translate frequency and independently route multiple harmonics remains limited when the coding state is static or quasistatic.

The review then explains the harmonic-generation mechanism of STCMs. A periodically varying reflection coefficient can be expanded into a Fourier series, and multiplication of this coefficient by the incident field produces components at the carrier frequency shifted by integer multiples of the modulation frequency. The complex Fourier coefficient of each order determines that harmonic's amplitude and initial phase. Duty ratio, switching phase, and time delay can therefore be designed to enhance or suppress selected harmonics and to control their phases. When the temporal sequence also varies across the aperture, spatial and temporal phase gradients jointly determine frequency and propagation direction. This framework connects electromagnetic coding with frequency-shift keying (FSK), phase-shift keying (PSK), quadrature amplitude modulation (QAM), beam scanning, nonreciprocal transmission, and space-frequency channelization.

Communication demonstrations are organized by the physical dimensions being controlled. Space-frequency multiplexing uses different harmonics and beam directions to carry independent data streams. Space-frequency-polarization multiplexing adds orthogonal polarization channels and has demonstrated parallel quadrature-phase-shift-keying (QPSK) transmission with a bit-error rate below 3 × 10⁻⁴ in the reviewed experiment. Near-field focusing creates several spatially separated communication spots from one programmable aperture. Duty-ratio and time-delay coding provide approximately independent amplitude and phase control for high-order millimeter-wave modulation, including 256-QAM. Other reported functions include light-to-microwave conversion controlled by visible-light waveforms, sideband-suppressed leaky-wave antennas, asynchronous multi-partition control, integrated radiation and stealth, and trackable electronic deception. Together, these examples show how the harmonic spectrum can serve as a set of programmable communication channels rather than as an unwanted by-product of switching.

The review further examines orbital angular momentum (OAM) and spatiotemporal vortex beams. A helical spatial phase can be combined with time coding so that selected harmonics carry prescribed topological charges. Experiments have demonstrated dynamic switching of OAM states, polarization-locked vortex generation, controlled superposition of several OAM modes, and joint use of OAM, frequency, and polarization for high-dimensional multiplexing. A multifunctional holographic STCM has also reported an OAM-mode purity of approximately 0.83 for a selected harmonic. These results illustrate the hardware flexibility of programmable apertures, but they do not imply unlimited capacity. Finite aperture size, beam divergence, imperfect sampling, phase quantization, misalignment, multipath propagation, and receiver dimensions reduce modal orthogonality and increase crosstalk. OAM performance should therefore be evaluated with mode spectra, crosstalk matrices, bit-error rate, and capacity under the same aperture and channel constraints used for conventional multiple-input multiple-output (MIMO) systems.

Overview STCMs provide a unified hardware route for modulation, radiation, frequency translation, and multidimensional channel control. The reviewed results support their potential in multichannel links, millimeter-wave and terahertz front ends, near-field communication, reconfigurable intelligent surface (RIS)-assisted networks, and integrated sensing and communication (ISAC). Their practical value will depend on progress beyond proof-of-concept demonstrations. Tunable elements must combine low insertion loss, adequate power handling, fast switching, broad bandwidth, and manufacturing consistency. Large arrays require reliable bias networks, clock distribution, thermal management, calibration, and suppression of undesired sidebands. Communication algorithms must account for hardware nonlinearity, modulation delay, harmonic leakage, channel variation, and the control overhead needed to update coding states.

Future work should establish reproducible system-level benchmarks that report spectral efficiency, error performance, energy efficiency, out-of-band radiation, synchronization accuracy, and robustness to mobility and multipath. Joint electromagnetic and communication optimization can then select the aperture layout, temporal code, waveform, and receiver together rather than treating them as independent components. Machine-learning-assisted synthesis may reduce the cost of searching large coding spaces, provided that constraints from devices and channels are included in training and validation. At the network level, STCM control should be coordinated with channel estimation, beam training, user scheduling, and sensing tasks. Progress along these lines would clarify where spatiotemporal coding offers a measurable advantage over conventional phased arrays, mixers, and passive RIS architectures and would support its transition from laboratory prototypes to deployable wireless systems.