“Wow, look! It’s moving! It’s moving!”
That’s what I was shouting while showing the movie to my student and our professor. It was a movie of a piece of ceramic in an oven. Basically it depicts something black on a grey background, nothing to get that excited about. If you were to watch it you would probably say: “Yeah, it’s moving, aaand …?” But the cool thing to get excited about is the fact that this was a dried flat stripe of ceramic programmed with some code inside so that it would change shape during heating. First time this was done in with a ceramic material.
Indeed, instead of aiming for the perfect flat square tile that costs so much efforts to obtain, we really wanted to deform our ceramic. Usually when you make a tile, for example by spreading clay with a rolling pin like you would do for making the crust of a pie, the corners always lift up during drying. Even if you managed to keep the thickness of the tile perfectly homogeneous, the corners are twice more in contact with the air and will dry twice more quickly. And usually you don’t want that.
In our research project, we wanted to make a ceramic piece that would fold, bend or twist, not only during drying, but more importantly during the firing step that leads to the stiff strong and brittle final ceramic. In our jargon we call this firing step sintering. It means heating at elevated temperature without melting the sample, typically from 1000 to 1800°C. Having a ceramic element that can shape only when in the oven would be a unique way to make interlocked ceramics or chains or encapsulated pieces with no cutting or gluing! Honestly, it is also because it is beautiful to see it moving during the sintering. We got the chance to get the video by sending our precious samples in Italy where they have a special furnace with an infra-red camera so they can record even when the temperature is so high that everything looks red.
Pictures of edamame soya bean in wet and dry states.
The brilliant idea of our professor was to imitate the opening of plant seedpods. For example, if you take a fresh edamame soya bean, carefully separate the two parts of the seedpod, and let them dry, you will see that they will twist!
When they are in fresh, the beans are composed some stiff long fibers, named cellulose fibers, similar to what you find in trees, surrounded by a very wet matrix. But the orientation of those fibers is not random, as they form a bilayer with perpendicular orientations. During the drying of the bean, the fibers impede the shrinkage of the matrix along their long direction so that the shrinkage can only take place perpendicularly to the fibers. This is an anisotropic shrinkage: it shrinks only in one direction. Since the orientation of the fibers is not the same in both layers, one layer will want to shrink in one direction while the other in the perpendicular direction. This leads to the change in shape during drying.
Schematics of the bilayer structure in the Edamame seed pod, where stiff cellulose fibers are oriented perpendicularly in a swellable matrix.
Depending on the angle between the two layers, you can go from twisting to bending. Also if the change in shape happens in drying, it also happens when further hydrating, since swelling and shrinking are the same only in opposite directions. In both cases, it is the anisotropy that induce the shape change. I invite you to take a sheet of paper A4. In this sheet, the paper fibers (cellulose again) are aligned along the long direction. Borrow the UHU stick from your neighbor’s kid and glue two pieces of paper (for example old bills) with perpendicular angle between the fibers (basically one paper has its shortest dimension before you while the other has its longest). Put glue all over the paper, not only at the corner. Then cut a stripe into this bilayer, with the angle of your choice between your cut and the direction of the fibers in the first layer (you may use your credit card as a ruler). Already when the glue is drying, you should observe the expected change in shape. To make more intense, humidify the piece of paper with some water and let it dry again. Don’t put too much water as the two sheets might peel off. So what do you get? Twisting or bending? You can also vary the geometry. I quite like the squares, they work well.
Schematics of fiber alignment in bilayer structures to induce bending or twisting after drying.
Pictures of bending and twisting shapes made from paper.
To do the same in a ceramic, we need create anisotropy. To do so, we align stiff elements in one direction. Since one of our previous colleague developed a technique to align disc-shape particles, we used them instead of fibers. These particles are called platelets and are 0.250 µm thick and 10 µm in diameter. A human hair has roughly a diameter of 100 µm, which is 0.1 mm. So our platelets have a diameter ten time smaller than that of a human hair and a thickness one thousand times smaller! You very probably have some of these particles around you: they are typically used in car paints but also in cosmetics because they give a shiny metallic color.
To create the anisotropic shrinkage, we had to apply a trick to align the platelets vertically and all in the same direction. What our colleague did, was attaching nanoparticles (more than ten thousand smaller than the diameter of our hairs) of iron oxide to the surface of the platelets. These nanoparticles contain iron, which gets attracted to magnets. That way, our platelets also get attracted to magnets.
The attachment of the iron oxide particles to the surface of our platelets happened spontaneously in water. The chemistry of the platelets we use is aluminium oxide, also called alumina, which is one major component of porcelain (that makes it white). When you mix the platelets in water, the surface gets positively charged. This is linked to the atoms at the surface and their affinity with atoms in water. For other chemistries, it might get negatively charged or no charge. On the opposite, we purchased iron oxide nanoparticles that are made to get negatively charged when put in water. Plus and minus gets attracted, therefore the iron oxide nanoparticles spontaneously get attached to the surface of the aluminium oxide platelets. We just witness one example of a so-called Van der Waals interaction.
Schematics of the principle to make particles that are attracted to magnets
When put in a liquid, the new platelets now get attracted to a magnet but also align vertically when a magnet approaches from above. To fully reproduce the alignment of the fibers, the platelets need to be aligned all parallel to each other. Rotating magnetic fields can be used to do that. The set-up consists only in a magnet attached to a motor and rotating fast enough. When the magnet rotates too slowly, the platelets have the time to rotate along with the magnet. However, if you rotate too fast, it becomes difficult for the platelet to follow. During each rotation, the platelet has to move a lot of the fluid and this requires a lot of energy. Therefore, when the platelet cannot follow anymore, it aligns in the plane of rotation of the magnet, where the rotation has no influence any more. All the platelets do the same and they align in parallel.
Now that we have the “fiber-equivalent”, what about the matrix? To mimick the edamame bean, gelatin as a matrix is convenient. Warm gelatin at 55 °C is liquid so alignment of the platelets with the rotating magnet can take place. The particles can easily move to align in a liquid. Removing the heating and letting cool down the gelatin to room temperature, it gels. The platelet cannot move anymore and their alignment is maintained. Pouring another layer of warm gelatin containing the platelet and rotating the magnet along another direction leads to the formation of a bilayer that will twist or bend when drying or swelling, as we described for the piece of paper. Look the pictures in the reference called “Self-shaping composites with programmable bio-inspired microstructures” which is open access (see link below). Everybody I’ve shown these pictures refer to it later as the “twisting pasta”. That would indeed be a very elegant way to make shape-changing cooking noodles!
In order to reproduce this in a ceramic, we choose a matrix that is initially very liquid and that gels with time to maintain the platelet orientation. This is thus the same principle as for the gelatin. To form a homogeneous ceramic, the chemistry of the matrix is the same than that of the platelets: aluminium oxide. We are therefore making a suspension of aluminium oxide platelets in aluminium oxide smaller particles.
We have seen that in water, aluminium oxide is positively charged. This charge is higher in acidic liquids, for example at pH 4, like in coke. Dispersed in these conditions, all the aluminium oxide particles are highly positive and therefore do not like each other: the suspension of particles in acidic water is very liquid with no agglomeration. However, at a basic pH of 9 like in diluted caustic soda, aluminium oxide exhibits zero surface charge. There is no more repulsion and the particles aggregate, forming a very viscous gel.
The increase in pH can be triggered internally by adding urea, a molecule, to the suspension of particles and by using an enzyme, urease. Enzymes are big biological molecules that are able to accelerate reactions. Urea is a typical fertilizer that gets transformed in the soils into ammonium and carbon dioxide by bacterias possessing the enzyme urease. It is the same reaction of urea with urease that we exploit here. During the reaction, molecules of hydroxide, OH–, are generated are lead to high pH, just as in caustic soda (which is also called sodium hydroxide). Along with OH–, carbon dioxide is also liberated as well as some other ions NH4+ that do not play a major role here.
So, we add urea and urease in our suspension of aluminium oxide at pH acid around 4, thoroughly mix then spread the liquid on a substrate below a rotating magnet. After only 12 minutes, the enzymatic reaction has occurred and the suspension transformed from a liquid to a gel with aligned platelets because the pH has raised from 4 to 9. A second layer is then poured on top with another orientation of the platelets. After 12 more minutes, we obtain a bilayer structure that can be cut in stripes as for the previous systems.
Schematics of the enzymatic reaction between urea and urease that leads to an increase in pH and gelation of the aluminium oxide particles.
After drying, since there are so many particles in our system, very little anistropic shrinkage happens, therefore the bilayer remains flat. To induce the shrinkage, we need to fire the ceramic at elevated temperature, when the particles can start to fuse between each other. This takes place in a special furnace and is usually between 1000 °C to 1250 °C for clay potteries or stoneware and porcelains. For aluminium oxide, the sintering typically happens at much higher temperature, between 1300 and 1700 °C. This is a reason why do not see aluminium oxide cups: high temperature demands high energy therefore high cost. Also, sintering, the technical term for firing, is influenced by the particle size. It is a diffusion process, meaning that the atoms are moving from one particle to another, forming a neck. Particles linked together by many necks forms a solid ceramic. In our case, we sintered the bilayers at 1600 °C… and observed the expected change in shape! (see the images in the open source paper titled “Bio-inspired self-shaping ceramics”.)
Schematics of the sintering process between two ceramics by heat-induced diffusion of atoms.
It is incredible to see how Nature is smart and that simple alignment of fibers or particles, even if the chemistry is the same, can lead to such impressive change in shape! The beauty is that we can exploit the variability of our synthetic world to reproduce the same strategy in more complex and/or smarter systems, such as hard ceramics. Also, now that we have understood the phenomena and how to reproduce edamame twisting, we can play around and create other shapes. My student was very creative and combined twisting and bending in various directions. He also proved that this strategy can be used to encapsulate a ceramic inside a closed vessel without need for cutting and gluing. I am still hoping he will be able to make ceramic ships in quartz bottles! Without joking, the method could lead to applications in specific fields where high temperature, harsh environment and complicated shapes or change in shape is required.
Picture of a complex self-shaping aluminium oxide ceramic that resembles a man playing tennis.
Picture of a complex self-shaping aluminium oxide ceramic that resembles a spider.
Erb, R.M., Sander, J.S., Grisch, R. & Studart, A.R., Self-shaping composites with programmable bio-inspired microstructures, Nature communications (2013), 4, 1712. http://www.nature.com/articles/ncomms2666
Bargardi, F.L., Le Ferrand, H., Libanori, R. & Studart, A.R., Bio-inspired self-shaping ceramics, Nature communications (2016), 7, 13912. http://www.nature.com/articles/ncomms13912
The movie can be found as Supplementary Movie 1 at the bottom of the page at: http://www.nature.com/articles/ncomms13912#supplementary-information
Erb, R.M., Libanori, R., Rothfuchs, N. & Studart, A.R., Composites reinforced in three dimensions by using low magnetic fields, Science (2012), 355, 199. http://science.sciencemag.org/content/335/6065/199