3D printed microfluidic device for testing oral cancer drugs

Most oral cancer drug screens run at flow rates between 0.5 and 20 μL/min, so channel cross-sections stay small. A 3D printed microfluidic device built by material jetting or stereolithography holds a rectangular channel down to roughly 200 μm wide by 100 μm tall. Below that, the printer's own resolution starts to round the corners and narrow the lumen in ways your flow model will not predict.

Aspect ratio matters more than absolute size. Keep width-to-height between 1:1 and 5:1 on the print bed. A 200 μm × 200 μm channel prints cleanly; a 2,000 μm × 100 μm slot tends to sag in the middle on a resin printer unless you add supports that are hard to remove without scratching the floor.

Wall thickness should be at least 1.5 mm around any perfused channel. Thin walls flex when you push a syringe, and a flexed wall changes the shear stress the cells see. For a device that clamps into a holder, 3 mm walls give you a stable reference surface.

Design the inlet and outlet as vertical ports with a flat boss at least 4 mm across. Tubing adapters seat on that boss, and a flat land means a compression fitting seals without glue. Chamfer the port edge 0.3 mm to stop the punch-through that happens when a blunt needle is pushed into a printed bore.

• 1
  Minimum channel width200 μm for reliable resin printing; 100 μm only on high-end systems
• 2
  Aspect ratioStay between 1:1 and 5:1 to avoid roof sag
• 3
  Wall thickness1.5 mm minimum, 3 mm where the device is clamped
• 4
  Port bossFlat land ≥4 mm across, 0.3 mm chamfer on the bore edge

The channel floor is the cell culture surface, so its chemistry decides whether your oral cancer cells attach. Standard acrylate stereolithography resin cures with residual monomers that leach into the medium and shift viability readings. For anything that touches cells, use a biocompatible resin rated for short-term contact, or print a mold and cast the flow layer in PDMS.

PDMS remains the reference material for oxygen permeability and optical clarity. A 3D printed microfluidic device still earns its place as the master: print the channel geometry once, cast PDMS against it, and you skip the SU-8 photolithography step. Surface roughness of the printed master transfers to the PDMS, so polish or vapor-smooth the master if you need an optical floor.

For manifolds, holders, and fluidic interconnects that never see cells, engineering thermoplastics are cheaper and tougher. ABS and PC print on FDM machines and hold threaded ports better than resin. PEEK and POM are the right call when the part sees solvents such as DMSO at working concentration.

Leaching is the risk you cannot see. Run a blank perfusion with culture medium for 24 hours, then check the medium for pH shift and cytotoxicity before you trust a printed channel with live cells.

• 1
  Cell-contact layerBiocompatible resin or cast PDMS
• 2
  Printed masterRigid resin; polish if optical clarity is needed
• 3
  Manifold and holderABS, PC, POM, or PEEK for solvent exposure

A printed channel is open on one side until you close it. Thermal bonding works for thermoplastics but collapses fine channels if the platens are not flat within a few micrometers. Solvent bonding gives a strong joint and a partly dissolved channel wall, which changes the cross-section near the bond line.

Adhesive bonding is the practical default for resin parts. Spin or roll a thin layer of medical-grade epoxy on the flat land, align the cover under a stereo microscope, and cure under weight. Keep the bond line under 50 μm or the adhesive squeezes into the channel and blocks it.

Pressure testing tells you whether the bond held. Fill the device with water, cap the outlet, and pressurize to 2 bar for 60 seconds. A syringe pump with a pressure sensor is enough. Any device that loses pressure gets rejected before it reaches a cell culture hood.

For a device that runs drug combinations, add a second check: dye perfusion through every channel path. Air bubbles and partial blockages show up immediately as uneven color fronts, and both ruin a viability assay.

Printing wins while the channel layout is still moving. A geometry change costs nothing but a new build file, so you can test five manifold layouts in a week. That is the real value of a 3D printed microfluidic device in an oral cancer drug study: fast iteration on fluidic design before the biology is locked.

Machining wins once the geometry settles and the material has to survive. PMMA and PC mill to Ra 0.2–0.8 μm, which gives a clear optical path for imaging and a surface cells attach to predictably. PEEK machines well and tolerates the solvents used to dissolve poorly soluble drug candidates.

A hybrid route is common. Machine the manifold block that carries the threaded ports and the clamping features, then print or cast the flow layer. The manifold needs flatness and thread strength; the flow layer needs fine features and fast changes.

GreatLight runs 127 high-precision CNC machines, including 16 simultaneous 5-axis centers, and holds ±0.005 mm on metal and plastic parts. For microfluidic manifolds and holders, that tolerance keeps port spacing and mating faces aligned across a batch. We also run custom 3D printing for the flow layer, so both halves come from one quote.

Send the channel drawing and the port layout. We return a DFM note within 12 hours covering minimum feature size, wall thickness, and bond land, plus a quote for whichever route fits your run size.