Login / Signup

Deconvolution volumetric additive manufacturing.

Antony OrthDaniel WebberYujie ZhangKathleen L SampsonHendrick W de HaanThomas LacelleRene LamDaphene Marques SolisShyamaleeswari DayanandanTaylor WaddellTasha LewisHayden K TaylorJonathan BoisvertChantal Paquet
Published in: Nature communications (2023)
Volumetric additive manufacturing techniques are a promising pathway to ultra-rapid light-based 3D fabrication. Their widespread adoption, however, demands significant improvement in print fidelity. Currently, volumetric additive manufacturing prints suffer from systematic undercuring of fine features, making it impossible to print objects containing a wide range of feature sizes, precluding effective adoption in many applications. Here, we uncover the reason for this limitation: light dose spread in the resin due to chemical diffusion and optical blurring, which becomes significant for features ⪅0.5 mm. We develop a model that quantitatively predicts the variation of print time with feature size and demonstrate a deconvolution method to correct for this error. This enables prints previously beyond the capabilities of volumetric additive manufacturing, such as a complex gyroid structure with variable thickness and a fine-toothed gear. These results position volumetric additive manufacturing as a mature 3D printing method, all but eliminating the gap to industry-standard print fidelity.
Keyphrases
  • machine learning
  • air pollution
  • high resolution
  • deep learning
  • optical coherence tomography
  • quantum dots
  • neural network