The process of turning our BC into a box
Bacterial cellulose (BC) has several useful and unique properties, but in its natural state, it does not satisfy all our criteria for a packaging solution. Namely, raw, unmodified BC isn’t strong enough on its own. Though the incorporation of polyhydroxybutyrate (PHB) granules into the BC helps to diminish some of these challenges, it alone would be insufficient in producing an ideal packaging material. The BioSculpting workflow was created to develop and iterate a BC packaging that overcomes many of BC’s shortcomings while meeting the needs of our end users.
In order to create an effective solution, we first defined our desired specifications using industry contacts. From our meetings with Chris Clark, Category Director of Star Produce, we began to build a list of requirements for our packaging, affectionately named the “C-list”:
Consumers shop with their eyes, and thus, the physical appearance of the packaging must be carefully considered. Properties like clarity are important for consumers as it allows them to see exactly what they are purchasing - this is especially crucial for time-sensitive commodities like fruit. Colour is also important as it allows for the development of visually appealing packaging. This helps to mask the naturally unpleasant colour of BC, which might otherwise discourage people from using the packaging. The addition of colour also aids in hiding cosmetic flaws on the fruit, which may turn away consumers from purchasing edible but imperfect fruits. More information on these criteria can be found on the colouring wiki page!
Our packaging needs to be no larger than conventional, plastic-based packaging. The dimensions and shape of our material should be similar to existing packaging in order to fit readily within existing shipping infrastructure. By ensuring that these properties are met, we are able to integrate ourselves effectively into existing packaging workflows; the more difficult it is to use our BC packaging on an industrial level, the less likely it is to be adopted by fruit packaging suppliers.
Fruit packaging has immense variability, with various sizes, materials, and properties necessary for different fruits. In order to focus our efforts over the summer and avoid the conflicting demands that come with a catch-all packaging solution, Chris recommended that we aim to replace a single type of plastic packaging. We chose to replace the plastic clamshell packaging commonly used for berries and other small produce. This is due to the current lack of recyclable alternatives in the market, allowing our project to differentiate itself from existing options while increasing the potential impact that widespread adoption might bring.
One of our self-imposed requirements was complete compostability on both an industrial and household level. Many alternative packaging materials already boast biodegradability, but oftentimes, this requires timescales that are infeasible on the personal or municipal level – in essence, many existing alternatives meet guidelines for being considered “biodegradable” but do not fully decompose within a single industrial composting cycle. Furthermore, many biodegradable plastics leave microplastics when composting. Thus, in order to avoid both of these issues, we opted to ensure our product could be fully compostable. More information on our efforts can be found on the compostability wiki page!
Before any prototyping could begin, we first had to explore previous work regarding the shaping and working of bacterial cellulose. Much of the literature work favoured creating and testing small samples of BC, neglecting to explore its functionality as a large-scale biomaterial. As such, some of our early inspiration came from the works of designers and artists who explored the usage of BC as a way to integrate sustainable biotechnology large-scale production in other industries. Notably, Jannis Hülsen’s Xylinum Cones (1, 2) and Xylinum Stool (3) and Suzanne Lee’s BioCouture (4) helped to establish a framework of knowledge from which to build the BioSculpting workflow off of, leading to the development of our first prototype.
Still, before we could start building our BC boxes, we had to make a decision regarding our method of construction. We could choose between additive manufacturing, through either modifying a 3D printer to accept BC-based composite filaments (5, 6) or through the layering of BC in our desired shape by controlling the air-media interface across multiple layers, or subtractive manufacturing, giving us the option to choose between creating large, 3D shapes and removing BC or sculpting BC sheets over a mould and trimming any excess. Each of these had advantages and disadvantages which are listed in the table below:
If not already obvious by the name of the subproject, our work over the summer was focussed on the development of mould-centred biological sculpting, which later became known as simply BioSculpting.
Our first prototype served as a proof of concept regarding the feasibility of our mould-centred approach. Though it has been done in the past, seasoned lab members know that oftentimes, getting something that should work to work is half the battle. In order to conserve BC in the event of a failure, our first prototype had an incredibly small objective: to produce a 4 cm x 4 cm-base box from scrap pieces of BC and to observe its properties, both wet and dry.
To begin, we first created a set of moulds to layer the BC over. These were simple hollow cubic shapes, with 0.5 cm thick walls surrounding a hollow interior. These moulds were created in sizes ranging from 2 cm wide at the smallest to 6 cm wide at the largest and were 3D printed using PLA.
Following this, we began proper work on creating our first prototype. We took several pieces of purified, treated BC, squeezed out as much water as possible from them, and layered the damp pieces onto the 4 cm x 4 cm x 4 cm mould. Where possible, we ensured pieces overlapped in order to ensure that they would merge in the drying process. When that was not possible, smaller pieces of BC were layered over the seams, serving as a “tape” for the mould. After layering was complete, the mould was put into a convection oven at 70°C for ~3h, before being left to air dry overnight.
Upon returning the next day, we had discovered that our prototype had largely succeeded. Despite substantial shrinkage in the z-direction, the resultant box had several notable properties. For one, the box itself was very strong considering its size, and was able to hold in excess of 120g before testing stopped due to size constraints. The box was also relatively elastic, able to withstand impacts and crushing without permanently compromising its shape. These properties inspired us to continue onwards with the BioSculpting approach, and soon after, development on our second prototype began.
Figure 1. The moulds used in the creation of prototype one. A variety of moulds were produced to allow for flexability depending on the amount of BC available to us.
Figure 2. Prototype one undergoing the drying process. The sample would be the first to undergo hybrid oven-air drying, a practice abandoned by our final prototypes.
Figure 3. Prototype one being loaded with mass to test it's strength. Note the shrinkage and warping present on the vertical faces of the box.
Following our initial successes, prototype 2 was born. Before we could begin creation, however, we first had to address the shrinkage issues seen in the first prototype; losing the majority of your packaging’s height is not exactly an ideal situation. After a brief investigative period, a hypothesis was formed: because BC loses the vast majority of its volume when drying, the BC “shrinks” and contracts onto the mould. However, the sides of the box were free-hanging, with only gravity (failing to) counter their slow, upwards contraction. With the hypothesis, a probable cause was found, and yet, nothing was done regarding the contraction. This was due to fears of overconstraining the system. If we had implemented a way for the system to restrict the shrinkage (i.e, folding the BC under the box to restrict it’s upwards movement), the BC would suffer internal stresses and strains as it contracted, potentially tearing the box apart. For a prototype in this stage, the shrinkage was a small price to pay to avoid a potential catastrophic failure. Thus, the shrinkage remained untouched for this second prototype.
The development of the second prototype followed quite logically from the first: we went bigger. In this iteration, we went from a 4 cm x 4 cm-base box to a 6 cm x 6 cm one. This allowed us to test two major questions. First, does the BioSculpting workflow scale across larger and larger prototypes? This was important to find out early as scalability issues compromise our ability to create an effective industry solution. Secondly, we needed to see if the BC retained its properties in larger form factors, as this would influence future design decisions.
The specific production workflow was near-identical to prototype one, layering pieces of BC onto the larger mould and letting it dry. The results, however, proved to be quite different from the original prototype. For one, this BC was significantly less elastic and more brittle, likely due to there being less BC/cm² compared to the initial prototype. This difference also manifested itself in a translucent, plastic-like film on some sections of the bottom face. This film, while possessing excellent visual qualities, was later revealed to have weaker physical properties than the other BC samples. The box, much like the first prototype, was able to hold a considerable amount of weight despite its small size. In testing, we found it was able to hold a combined 436.24 g of weight (despite weighing only 0.9 g) before we ran out of weights. These results further incentivized us to pursue BioSculpting as a viable strategy to produce our packaging, but in order to combat the relative thinness of the first two prototypes, a new tactic had to be employed.
Figure 4. Prototype two prior to the drying process, featuring more uniformity in BC thickness.
Figure 5. The dried prototype two being help up to the light to showcase its transparency.
Figure 6. Weight testing of prototype two. Batteries were added to the weighing due to us exhausting our quarter supply.
In order to combat some of the shortcomings of our first prototypes, we began work on finding a solution to enhance the strength and thickness of our second generation of BC packaging. For inspiration, we decided to look at what has been done before with BC’s compositional brother, plant cellulose. Plant cellulose has been used by humanity in a variety of fields for centuries, and packaging is no exception. Paper, a thin, cellulose-based material, is used to create one of the most versatile packaging materials in existence: cardboard. Cardboard’s strength comes from its ingenious multilayer design; it is simply two flat pieces of paper (called liners) with a wavy piece (called the flute) glued in the middle. The flute serves as a flexible cushion, allowing for stronger properties than could be achieved by stacking flat sheets of paper alone. It is this strength that we aimed to replicate and emulate in prototype three.
Our work on this prototype began with the creation of a new set of moulds. First, we needed a way to create the wavy, central flute for the cardboard. Traditional cardboard manufacturing uses a “fluting machine” to press the cardboard paper into the wave-like shape. Our mould utilises a similar technique, just on a much smaller and less time-efficient scale. We modelled and 3D printed two rack gears with the intention of sandwiching the BC in between them as it dried. To replicate the downward forces that fluting machines use, weights (in the form of test tube heating blocks) were added to the top of the moulds, allowing the moulds to keep pushing down into the BC even as it dried and lost volume. Once the BC had dried fully, it was removed from the mould and had retained its wave-like shape.
Figure 7. Wet BC being pressed into the corrugated shape as it dries.
Figure 8. Top view of the corrugated piece, showcasing the uniformity in the corrugation.
Figure 9. Side view of the corrugated piece, showcasing the amplitude/depth of the waves.
The second set of moulds were much simpler than the last: we 3D printed two flat moulds with the same profile as our rack gears to create the liners necessary for cardboard manufacturing. Relative to the novelness of the flute, the liners provided a return to normalcy, with no exceptional methods or tactics employed in their creation.
Producing the cardboard from the three separate pieces involved exploiting a useful property of BC. Normally, cardboard uses a strong glue to adhere the flute to the liners. However, in the interest of maintaining full biodegradability, we decided to use water instead of a conventional adhesive. BC samples that are stuck together when wet remain attached upon drying. This “water glue” allowed us to manufacture the cardboard with relative ease. We simply wet the liner pieces and pressed them onto the flute. After letting them dry overnight, we returned to a single, cohesive piece of BC cardboard. Qualitatively, the cardboard was flexible, relatively elastic, and had excellent compressive qualities. However, due to our lack of measuring equipment, many quantitative measurements used for regular cardboard (such as bursting strength and compressive strength) were not able to be found.
After creating our BC cardboard proof of concept, we moved onto creating a corrugated box prototype. This allowed us to see if BC cardboard could be scaled up, while also allowing us to see its impact on a larger, more practical scale. To do so, we first began by designing and printing a new set of moulds with two main changes. First, the box was designed using Fusion 360’s sheet metal function as a way to prototype foldable BC boxes. Because of its ability to create flat patterns from the three-dimensional box shape, it was relatively easy to export the 2D map of the structure to serve as the base for our mould. Our second change was the standardisation of the teeth dimensions on the mould itself. While the first moulds were created from relatively arbitrary dimensions, this second generation based its measurements on the C-flute sizing (7) in order to more accurately compare our BC cardboard with existing packaging.
After producing the moulds, the box was produced in similar fashion to the BC cardboard, with the three layers produced separately before ultimately being combined using our water glueing technique. The prototype produced was a strong, durable box, emulating the cardboard it was inspired by. Much like the first prototype, the added thickness allowed for the box to maintain a good balance of malleability and rigidity, allowing it to remain flexible and not excessively brittle. For the same reasons as the BC cardboard, we were unfortunately unable to get the quantitative measurements that cardboard boxes are usually compared by.
Figure 10. Prototype three, showcasing a small scale corrugated box prototype.
Figure 11. View of prototype three, allowing visability of the individual layers that make up the walls.
Our final prototype involved scaling up the methods developed in the third prototype. With the last prototype, we had proven that BC cardboard can be used to create a box. However, materials may behave differently when scaled up. We aimed to prove that our box could function at a scale seen in traditional packaging. We chose to replicate 6 oz (170g) raspberry clamshell packaging (8), as it allowed us to test our box against our direct competitors, while also offering a goal that was achievable in the timeframe available.
The box was produced using the same workflow as above. One of the key differences in design was the presence of a lid piece, fully enclosing the box. The box was designed to be 12 cm x 12 cm x 4.5 cm, occupying 720cm³ of space. We created two-dimensional nets of our box, developed 3D-printed moulds, and proceeded to lay our BC onto it. This created individual sheets of material that were then pressed into BC cardboard. The material was then folded into the box shape and left to dry.
The box produced was lightweight relative to its size and its strength. It retained the flexibility seen in prototype three, and featured exceptional tear resistance. The sides and top were malleable, similar to cardboard, and were able to bend without failure. The bottom, however, retained the stiffness and strength necessary to support the 170g of weight we hoped to insert on it. However, due to the lack of appropriate equipment and the desire to keep the box intact as a visual aid for the Grand Jamboree, we were unable to quantitatively measure the strength of the box, nor test it to failure.
Figure 12. Top view of the prototype four liner, left out to air dry.
Figure 13. Creation of the corrugated component of our final prototype.
Figure 14. Final box created after folding up the 3D net, left to air dry overnight.
In the short term, our future directions will primarily focus on the creation of increasingly larger prototypes. This allows us to test the workflow on a larger scale and create alternatives for different sizes and shapes of clamshell packaging. Thinking further ahead, we want to explore the possibility of hybrid forms of packaging, incorporating hydrogen-peroxide-purified films and windows into our product to allow for greater variety in the packaging produced. This could allow us to expand beyond just clamshells, replacing plastics in other products and fruit packagings.
Finally, we plan to further iterate the development of BC cardboard production and develop more effective, efficient methods of production in order to better compete with the rapid production used for other packaging materials. While the time-to-produce is a large goal, we also don’t want to neglect our desire to remain sustainable. We want to explore methods to reduce the amount of BC wasted in the production of our packaging, find alternative uses for scrap pieces such as the production of composite BC sheets (like particle board or oriented strand board for wood), and, if not possible, incorporate wasted material into the composting workflows developed in other parts of the project.
The files used to produce the moulds used in the subproject can be found here!