Answers are grounded in this paper and cite their sources.
Full field lithography shows potential to improve throughput in oversized packages, suggesting advancements for AI applications.
The performance specifications of artificial intelligence (AI) hardware drives the requirements for electronic systems to new levels. In order to deliver on the extremely high data transfer rates, low latencies, and escalating power requirements inherent to AI applications, significantly higher level of integration between logic, memory, and power supply components must be achieved than in traditional computer architectures. Heterogeneous packaging marks a pivotal development in the field of Advanced Packaging for electronic components for these applications. A key trend shaping this landscape is the shift toward significantly larger package sizes to integrate the divers chips. Modern heterogeneous packages frequently have a size of multiple exposure fields of a stepper, with designs approaching full wafer dimensions now under consideration. These oversized packages introduce substantial challenges for existing lithography systems. The expanded scale complicates stepper exposure processes, necessitating stitching of multiple step-fields and increasing susceptibility to within-package warpage and topographical irregularities caused by varying chip heights. Moreover, the use of extremely thick resist layers – often needed for fabricating through-substrate vias – further strains lithographic precision and resolution. Conventional stepper-based lithography, which relies on small-field reticles, is ill-suited to these large-area substrates. Stitching multiple exposure fields introduces overlay errors, process inefficiencies, and severe limitations in yield and throughput. To address these limitations, full-field exposure tools – capable of exposing an entire package area in a single shot – are gaining attention as a compelling solution.
No takes yet. Share an insight, caveat, or question.
Hennemeyer et al. (2025) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: