Underwater 3D printing of concrete: Mapping mix-design drivers of hardened performances

Underwater 3D concrete printing (UW3DCP) exposes fresh layers to water films and, at times, flow, making hardened-state gas permeability a decisive descriptor of the printed microstructure. This study quantifies the influence of mix design and placement conditions on intrinsic permeability k int for specimens cast and printed, in air and underwater, and in two directions (along filaments L, across layers T). Four sand grading upper-size bounds (0–0.5–0–2.0 mm) are investigated. Two rheology modifiers, cellulose ether and a Sphingan gum, are used systematically. Intrinsic permeability clearly separates both process and medium at low anti-washout dosage: printing and immersion increase k int , with the T direction the most affected. Increasing Sphingan gum content forms a high-strength micro-gel that reduces the level and temporal evolution of k int through the deposition-to-set window, and also reduces permeability anisotropy. The role of an imposed underwater current is quantified to isolate forced washout. Once paste rheology is held fixed, the grading upper bound becomes a first-order control: biopolymer chemistry alone cannot compensate for coarse fractions. Cellulose ethers primarily stabilizes air, and has limited leverage on hardened k int . Sphingan gum reduces k int without increasing density. Overall, k int is shown to be a robust acceptance parameter for UW3DCP, sensitive to seam formation, flow exposure and aggregate grading. The results provide design guidance: cellulose ethers should be minimized to the pumpability threshold; Sphingan gum should be dosed to neutralize current effects; and sand grading upper-size should be limited to improve packing. • Intrinsic gas permeability is established as an indicator of microstructural quality and anisotropy in Under Water 3D Concrete Printing (UW3DCP). • Sphingan gum is highly beneficial to effectively suppresses both temporal drift and directional anisotropy of permeability. • Maximum particle size remains first-order: biopolymers cannot fully compensate coarse grading. • Sphingan-cellulose blend delivers the optimal pumpability/washout trade-off and outperforms cellulose ether alone.