T. Ching*, S.-Y. Chang*, T. Takenouchi, Y. S. Zhang, Y.-C. Toh, M. Hashimoto
*Co-first author
Small Methods, 10.4(2026) e00416. [Paper Link]
Most organ-on-chip devices are still made by soft lithography. Every design change means a new mask, a new mold, and days of waiting. Once cells are in the chip, the design is fixed. But biology experiments rarely go as planned. Often you only know what the chip should look like after you see how the cells behave. Systems that can pattern inside a chip in real time do exist, but they rely on expensive optics. I wanted a way to change a chip's design as quickly as changing a picture on a screen.
The system uses a low-cost LCD screen (under $100) as a photomask that can be redrawn at any time. A chip filled with a light-curable hydrogel sits on the screen. A red light and a cheap digital microscope (under $50) let me see the chip and line up the pattern without curing anything. Then a blue light cures the hydrogel only where the screen lets light through, forming the channels inside the chip. All built and controlled using Arduino.
An LCD screen is made of a grid of tiny pixels, in our setup, each about 32 µm wide. Each pixel can be switched on or off. When a pixel is on, light from below passes through it. When it's off, the light is blocked. By displaying an image on the screen, I control exactly where light reaches the chip above it. The hydrogel cures only where light passes through, so the image on the screen becomes the shape of the channels. Changing the design is as simple as changing the image. Pixels can also be set to shades of gray, letting through partial light to control how tall the cured hydrogel grows.
Because the mask is just an image, changing a design takes minutes instead of days. I made channel networks inspired by capillaries, liver lobules, and the kidney glomerulus, each by simply editing the image. A full design change, from edit to patterned chip, took under 5 minutes. The smallest perfusable channels were about 65 µm wide.
Beads flowing though the vasculature network
Because the camera shows where things are in real time, I could pattern around cells that were already in the chip. I trapped liver cancer spheroids by anchoring, embedding, or caging them with hydrogel, then perfused them for 5 days. Caged spheroids showed almost no cell death, while spheroids in static culture died in the core. The same approach let me pattern multiple cell types side by side and build endothelial barriers that model cancer cells escaping into blood vessels.
Read the full paper in Small Methods: [link to DOI 10.1002/smtd.202400416].