Objective On-chip optical phased arrays (OPAs) provide compact structure, high integration, and precise beam steering capability. These features support their application in free-space optical communication (FSOC) systems. Single-aperture OPAs suffer from limited emission power and relatively large beam divergence. Multi-aperture cooperative transmission improves emission power and far-field directivity. Optical path differences exist between apertures in such structures. Relative time delay and piston phase difference are generated between emitted signals. Time delay causes pulse broadening at the receiver and induces intersymbol interference (ISI). Piston phase difference reduces far-field power concentration, defined by power in the bucket (PIB), and leads to a decrease in received signal-to-noise ratio (SNR). Communication performance is degraded by the combined effects of these two factors. Existing studies focus on coherent beam combining, while communication performance analysis remains insufficient. Quantitative evaluation of time delay and piston phase difference effects is required for system design.

Methods A joint simulation framework was established using OptiSystem and MATLAB. A cooperative communication system based on on-chip multi-aperture OPAs was constructed. A 10 Gbps NRZ-OOK modulation format was used. The carrier wavelength was 1550 nm. The receiver employed an intensity modulation and direct detection scheme. A photodetector and a low-pass filter with 10 GHz bandwidth were adopted. Controllable time delay was introduced in different aperture branches. Two-aperture and six-aperture configurations were analyzed. Bit error rate (BER), eye diagrams, and signal waveforms were used to evaluate communication performance. A near-field complex amplitude model was built in MATLAB for piston phase difference analysis. Far-field intensity distribution was obtained using Fourier transform. Power in the bucket (PIB) was used to characterize far-field energy concentration. Variations in PIB were mapped to received SNR. These SNR values were applied in the OptiSystem model to obtain BER results.

Combined effects of time delay and piston phase difference were evaluated by introducing both impairments simultaneously. An experimental system with two apertures was constructed. Relative time delay was adjusted using an optical fiber delay line. A lithium niobate phase modulator was controlled by a stochastic parallel gradient descent algorithm to achieve co-phasing control.

Results and Discussions Time delay shows a clear impact on communication performance in the two-aperture configuration. When the delay is below approximately 60 ps, ISI remains weak and BER stays at a low level. BER exceeds 10⁻⁹ when the delay reaches approximately 70 ps. Communication requirements are no longer satisfied at this point. Time delay in the six-aperture system follows a random distribution. BER remains below 10⁻⁹ when the delay range does not exceed 80 ps. BER increases significantly when the range exceeds 80 ps. System stability is degraded.Piston phase difference reduces normalized far-field PIB from 0.74 to approximately 0.31. Received signal power decreases, SNR decreases accordingly. BER increases with increasing phase difference. Under zero time delay, the phase difference tolerance is approximately 0.455 π for BER below 10⁻⁹. Combined effects of time delay and piston phase difference further degrade system performance. Tolerance ranges are reduced. Under a time delay of 50 ps, phase difference tolerance decreases to approximately 0.262 π. Time delay induces ISI and enhances performance degradation at a given phase difference. Phase difference reduces SNR and lowers tolerance to time delay. These effects jointly compress the system performance margin.

Experimental results are close to simulation results and show consistent variation trends. Under co-phasing control, BER starts to increase at a time delay of 60 ps. BER exceeds 10⁻⁹ when the delay is greater than 75 ps. Independent compensation methods were also evaluated. Co-phasing control alone or time-delay compensation alone improves performance only partially. Simultaneous application of both methods maintains error-free performance.

Conclusions Time delay and piston phase difference affect NRZ-OOK communication performance in on-chip multi-aperture OPA systems. Time delay tolerance is approximately 70 ps in the two-aperture configuration. Phase difference tolerance is approximately 0.455 π under zero delay. Delay range should be controlled within 80 ps in the six-aperture configuration to maintain BER below 10⁻⁹. Combined impairments reduce system tolerance and compress performance margin. Co-phasing control and time-delay compensation are both required for stable communication.Simulation results are supported by experimental observations. These results provide quantitative guidance for parameter design and synchronization control in on-chip multi-aperture OPA communication systems.