From Glass Microchannels to Plumbing
Research notebook · September 26, 2026 · Demonstrated microscopic channels; proposed building services.
The appealing idea is simple: make the route the water needs, already inside the object that surrounds it. No elbow fitting where a gentle curve would do. No concealed coupling merely because a worker needed a way to assemble the pipe.
Glass research gives us a real starting point. It also makes the distance to household plumbing wonderfully concrete.
Two demonstrations worth distinguishing
In 2017, Kotz and colleagues reported three-dimensional printing of transparent fused silica glass. They shaped a photocurable silica nanocomposite and used thermal processing to obtain glass. The starting feedstock and the final glass were different stages of the process; this was not a construction robot squirting finished glass pipes into a wall.
Their group’s 2019 hollow-microstructure paper used a different idea: make a sacrificial polymer template, surround it with glass-forming material, then remove the template during processing. The void preserves the channel’s shape. Some templates were additively manufactured. The resulting fused-silica objects contained complex microscopic passages, with substantial shrinkage accounted for during conversion.
The important distinction is that the second demonstration combined template manufacture, molding, and thermal processing. Calling it simply “printed plumbing” would hide most of the engineering. Its achievement is complex enclosed glass geometry at a small scale, not a qualified drinking-water system.
What changes when we make it bigger?
For the house project, channel diameter is only one variable. Long runs have to drain during manufacture and remain inspectable afterward. Temporary material needs an exit. A thick surrounding structure changes the heating and cooling problem. Every branch has to come out where the digital model says it will, despite processing shrinkage.
Then we have the job of the finished channel. Supply water, hot water, wastewater, and a closed heating loop are different applications. Each needs its own choice of material and evidence about pressure, temperature, chemical exposure, cleaning, and compatibility with the surrounding structure. Smoothness alone does not establish resistance to biological growth or scaling.
An integrated channel also creates a repair question. If a defect develops, can the house isolate that section, inspect it, bypass it, or renew its lining? A continuous route with no assembly joints could still contain manufacturing defects. I want to eliminate avoidable failures without making the remaining ones inaccessible.
A useful first experiment
Start with a non-potable laboratory channel specimen, including a curve, a branch, and an inspection or cleaning opening. Choose one manufacturing route and make its entire sequence explicit. Measure its final geometry, surface condition, wall thickness, and defects before judging fluid performance.
Then test leakage, pressure cycling, thermal cycling, and cleaning access under a declared experimental envelope. Compare it with an assembled reference channel serving the same purpose. Record total processing time and energy, including template removal and finishing, rather than reporting only printer motion.
If that succeeds, embed the channel in a larger test panel and repeat the tests while the surrounding panel experiences realistic restraint and temperature differences. Eventually, testing for a specific building-water application would be a separate qualification step. We should be able to point to each rung of that ladder.
The first win need not be an entire bathroom. A repeatable curved passage with a tested interface and a workable recovery process would already be something another team could build on.
Related pages: ceramic processing and material transitions.