Additive manufacturing, also known as 3D printing, is an invaluable platform for fabricating complex structures that would otherwise be time-consuming, or even impossible, to make using conventional techniques. The ultimate goal is to fully control the composition, geometry and properties of 3D-printed structures at a microscopic level. However, for objects made from multiple materials, the minimum size of features that can be printed is limited, because the rate at which printers can switch between materials is too slow. New technology is therefore needed to print fully functional, multi-material devices using a single printer. Writing in Nature, Skylar-Scott et al.1 report a micrometre-scale printing strategy that enables fast switching between viscous materials extruded through a single nozzle. The authors show that they can reduce the total printing time of certain structures by arranging up to 128 nozzles in parallel — a game changer for the field.
The 3D printing of soft or biological materials is often achieved through a process called direct ink writing, which involves the extrusion of a pressurized, viscous fluid through a single, moving nozzle. To print multiple materials using this method, the standard approach is to switch mechanically between nozzles, but this limits the minimum switching time. Another strategy is to sequentially push several materials through a single nozzle2,3, but it has not been possible to produce sharp transitions between materials in this way, or to achieve switching rates above 1 hertz.
To solve these problems, Skylar-Scott et al. have developed a microfluidic nozzle that brings up to eight viscous fluids together as separate filaments just before the tip of the nozzle. They make smart use of the fact that their printing fluids flow only when the internal stresses are above a certain value. By sequentially pressurizing individual fluids, they can switch between materials at rates of up to 50 Hz, and produce features at a scale of approximately 250 micrometres.
These switching rates are high enough to print ‘voxelated’ structures — in which each point (voxel) in a 3D grid that represents the structure can have different material properties (Fig. 1). The printing of voxelated structures has so far been demonstrated only for low-viscosity fluids using inkjet-based methods4 (which involve the propulsion of droplets rather than the extrusion of filaments). Skylar-Scott and colleagues’ work expands the potential range of materials that can be printed at these small feature sizes, and thereby opens up a range of applications for 3D printing that require precise control of local material properties.
The authors demonstrated the effectiveness of their approach by printing two functional objects that have periodic layouts of voxels. The first was a Miura origami pattern5: a sheet consisting of tessellating parallelograms. Skylar-Scott and colleagues printed this from a stiff epoxy material, connected by folds made from a second epoxy that was approximately 1,000 times softer. The object can be reversibly transformed from a flat shape to a folded, compact state by manually applying a force (see Fig. 4f of the paper1).
The second object was a soft robot6 made from two forms of silicone rubber of different stiffnesses (see Fig. 5 of the paper1). The robot’s legs consist of chambers that deform in a predefined direction when deflated and inflated; sequential inflation and deflation therefore results in a walking motion. For both objects, the use of multiple parallel nozzles was instrumental in reducing the printing time. This was important because the printing fluids start to harden continuously once made, limiting the window for using them.
Skylar-Scott and colleagues’ multi-material, multi-nozzle technique could have major implications for the development of ‘architected’ materials7 — those that exhibit exotic properties arising from their engineered, periodic substructures rather than their chemistry. Examples include materials that are extremely light yet strong8, and materials whose mechanical, optical or acoustic properties can be tuned by reconfiguring their internal structures9. Most architected materials so far have been made from a single non-architected compound. The ability to control the make-up of objects at a microscopic level (by printing combinations of voxels of different substances) opens up a new playing field, in which more and innovative functionalities can be programmed into the same architected material. This might lead to the production of architected materials that exhibit more machine-like behaviour than is currently possible10.
But we are not there yet. The available library of printable materials, and the range of properties represented, needs to be extended — for example, to include materials that have a variety of electrical and thermal conductivities, or that swell when they absorb a solvent. Moreover, at present, the spacing between the nozzles in the multi-nozzle printheads is unchangeable, and all the nozzles eject fluid simultaneously and at the same rate. This means that Skylar-Scott and colleagues’ system speeds up printing only for periodic structures in which the spacing between the nozzles determines the size of the periodic components. A different multi-nozzle printhead will be needed to produce structures that have other periodicities.
If the spacing between the nozzles was increased, then an alternative application of the multi-nozzle system could be to print exact copies of the same object in parallel. Work will also be needed to increase the flexibility of the technology, by making it possible to independently program the flow through each nozzle in a printhead, as is the case for inkjet-based methods.
Skylar-Scott et al. have pushed the boundaries of achievable speed and materials in additive manufacturing technologies. The work brings us closer than ever to being able to control the composition, geometry and properties of structures so small that they cannot be seen by the naked eye. This breakthrough is not merely a practical advance: it will change the way we design, build and think about functional devices.
Nature 575, 289-290 (2019)