Tens of millions of people each year find themselves in need of replacement body parts, but face a shortage in donor supplies. Many tissue engineering approaches (i.e., growing replacement body parts) have made great progress but still are limited in the complexity of the shapes that can be grown. My work was some of the early attempts to 3D print living tissue. The idea was to 3D print a hydrogel filled with a patient’s own cells in complex shapes directly from CT/MRI images. I modified a hydrogel material that had been used in other tissue engineering techniques so that it would be compatible with 3D printing.
Turkey, scallops, celery. Put it in a syringe, load it in a 3D printer, and dinner is served. People often ask if it tasted good: yes, it did. Then again, any time French Culinary Institute chefs prepare the printing material it is bound to be delicious.
Using chemistry tricks to modify materials, you can print nearly anything. Back in 2002 when I started working at the Cornell Creative Machines Lab, 3D printing was almost exclusively used for making plastic design prototypes. Our lab made some of the earliest attempts to 3D print non-traditional, functional materials. We printed functional electrical batteries by depositing zinc chemical slurries. Later, the lab went on to print functional actuators, transistors, capacitors, and more. This work set the stage for all the other work on printing non-traditional materials, such as tissue and food (below).
Most 3D printers operate entirely open-loop. In other words, a computer tells them where to place material but the system doesn’t check that material was actually deposited. With non-traditional printing materials, often times the materials exhibit erratic behavior. In order to get acceptable printing results with these tricky materials, closed-loop printing approaches are necessary to compensate for a variety of naturally occurring errors during the printing process. I used lasers to monitor the shape of the part mid-print and developed algorithms for correcting shape errors during prints. The technique enabled printing parts with high geometric fidelity despite uncertainty in the material behavior and environmental conditions.
What if 3D printers understood the substrate onto which they were printing? Most printers simply print onto flat surfaces, and planning the deposition tool's path is simple because it is "painting a blank canvas." If printers, however, could fabricate directly onto complex geometries then they could repair objects, for example. For this demonstration of in-situ repair, i.e. in-place repair, I started by printed a robot body out of epoxy. I then broke off one of the legs. Finally, I developed a printing method whereby the 3D printer laser scanned the robot, compared it to an image of the "healthy" robot, and then printed back the difference directly onto the broken robot.