Following the death of Thomas Edison in 1931, Nikola Tesla was scathing of his erstwhile colleague’s scientific methodology: “It was inefficient in the extreme. An immense ground had to be covered to get anything at all – unless blind chance intervened.” Edison’s trial and error methods were anathema to Tesla, “I was a sorry witness, knowing that just a little theory and calculation would have saved him 90% of the labour”.

Tesla’s genius was unique. Whereas Edison was the meticulous experimenter and tinkerer, Tesla was a human calculator. His ability to work out complex maths and physics equations helped him achieve early career success in Europe. In the digital age, this approach has been taken to the extreme by machine learning, with computing power capable of processing such equations in a matter of seconds at times. Across swathes of heavy industry, pharmaceuticals, software design and even in arts and culture, AI has brought experimental efficiency and the slow eradication of systematic trial and error in the lab. The superexpanding field of biosynthetic materials and 3Dprintable bio-inks is no different. In fact, without machine learning, a new 3D-printing technique that is helping amass a vast library of materials for use in all industries.

The work of aerospace and mechanical engineer professor Yanliang Zhang and his team at the University of Notre Dame, Indiana, the US, has been a magnet for global attention in this field. His career is built around research into thermoelectric materials, but he has accrued knowledge about bio-inks and their flexibility for 3D printing. His most recent research contributions show how high-throughput combinatorial printing (HTCP) can aid in the quest for materials used in multidisciplinary contexts. Although Zhang has an aerospace engineering background, the tentacles of his groups’ work cut across many sectors.

Zhang has been working towards an autonomous system for designing mixes of multiple aerosolised nanomaterial inks drawn through a single printing nozzle. The “art”, he says, is in controlling the ink blend ratios rapidly during the printing process. “Before, we were looking at one or maybe two decades to discover a material with a valuable application, be it in engineering, medicine or other sectors,” Zhang adds. “With medical applications, we hope to reduce that timeline of discovery to less than a year, even down to a few months, if we can create a functioning autonomous and self-driving process for materials discovery and device manufacturing.”

Using HTCP, Zhang and his team can control both the printed materials’ 3D architecture and ink compositions. His group can now produce materials with “gradiated compositions” and properties at microscale spatial resolution. In medicine, the potential application for these materials is vast for use as protective coatings around joints and tendons and many as yet unexplored areas of orthopaedics.

The new alchemy

The momentum for this revolution in both engineering and chemistry stems from the urgency to develop inks and machines to handle the tolerances needed to create bespoke materials. The physical chemistry involved is fascinating too, with a smorgasbord of chemical complexes available. HTCP allows direct fabrication of thin films of metals, nitrides, carbides, chalcogenides and halides – materials that outwardly would seem to be incompatible. The method of screening combinatorial materials is opening doors to structures thought impossible only a few years ago. For example, Zhang’s team report that the aerosol jet printhead has nozzles engineered with variable sizes that can deliver spatial resolution as low as around 20­m in the x–y plane and a deposition thickness as low as approximately 100nm.

“We printed a polymeric scaffold using material laced with Sporosarcina pasteurii, a bacterium that in nature starts the process of mineralisation to calcium carbonate deposition. After four days, the bacteria triggered the mineralisation process in the scaffold, and we were handling a fi nal product with a mineral content of over 90%.”
Esther Amstad

In an optimised range of ink flow rates, Zhang and co-workers found that this monotonic trend can be applied to a variety of nanomaterial inks, including silver nanowires (AgNW), graphene, thermoelectric compounds like bismuth telluride and even polystyrene. Before HTCP, the trialand- error system required extensive processing time. It not only caused difficulty with highthroughput fabrication, but could lead to messy side reactions, in some cases via a mismatch of starting materials related to their surface charge, pH values or ionic strength. Zhang uses nanoparticles of MXene – a 2D inorganic compound with a similar structure to graphene – and antimony telluride (Sb2Te3), as an example of such a mismatch. “MXene and Sb2Te3 nanoparticles exhibit opposite surface charges in a certain pH range, [which] leads to the formation of larger aggregates with poor colloidal stability.” But HTCP, Zhang adds, enables the rapid fabrication of combinatorial “samples”, minimising these unwanted side effects and resulting in a stable material with advantages conveyed by both compounds.

Sustainable and resilient

In Europe, a laboratory in Switzerland is also pushing the technological envelope. Esther Amstad is head of the soft materials laboratory at the renowned Federal Institute of Technology. She too is making waves with her bio-ink studies by exploring which types of compounds make useful biomaterials that are also environmentally friendly. “The need to be able to fabricate more sustainable materials and to process them in an energy-efficient, benign way has motivated me to examine 3D-printing options,” she says. “Our technology is paving the way to customised products with minimal material waste. Much useful research has been conducted designing the printing technology, but less has been devoted to ink formulation.”

How does Amstad see the long-term prospects for rapidly growing an array or library of bio ink coatings? “It depends on what is meant by longterm,” she says. “Bio-inks we study here help us produce viable and elegant ‘scaffolds’ for tissue engineering. These enable the growth of more functional organoids. But we can use them to test the effect of a variety of active substances, for example in the food and pharma sectors.”

According to Amstad, the range of materials that can be 3D printed is still limited. “One reason is that inks must fulfil certain complex flow conditions,” she says. “They must behave as a solid when at rest, but still be extrudable through a 3D-printing nozzle – with a texture a bit like ketchup.” Small mineral particles that have previously been used to meet some of these flow criteria have resulted in structures that tend to be soft or shrink when drying, which leads to cracking and loss of control over the shape of the final product. “We turned to the natural world for a potential solution,” Amstad adds. “We printed a polymeric scaffold using material laced with Sporosarcina pasteurii, a bacterium that in nature starts the process of mineralisation to calcium carbonate deposition. After four days, the bacteria triggered the mineralisation process in the scaffold, and we were handling a final product with a mineral content of over 90%.” The result was a strong and resilient bio-composite that can be produced using a standard 3D printer and natural materials, without the extreme temperatures often required for manufacturing ceramics. Also, the final products no longer contain living bacteria as they are submerged in ethanol at the end of the mineralisation process.

Regenerative medicine

The aptly named “BactoInk” opens possibilities that cut across many sectors. For biomedical use in particular, Amstad explains that “with additional work and funding, this technology will be a significant stride for personalised medicine”. “It would be marvellous to enable the fabrication of more customisable wound healing and drug delivery systems, open up new possibilities for the real-time monitoring of patients. One day, they might also be used in regenerative medicine, because of their ability to attain customisable shapes and properties.”

Unlike in Zhang’s lab, the use of AI to turbocharge the creation of bio-compatible composites has yet to become viable in Amstad’s work. It is only a matter of time and funding, however. “I’m hoping that the addition of AI will lead soon to in-vivo trials using bio-ink coatings taking place in humans and animal models, rather the trials we perform using art statues,” she says.

The work to develop bio-inks and composites using theory and calculation follows the trajectory set by Nikola Tesla almost a century ago, and with machine learning driving the process, researchers no longer have to be a human calculator to reduce their time spent tinkering in the lab.