BioForms I technical workflow

In order to generate a direct dialogue between the fabricated skins and mycelium, the form was generated to encourage mycelium growth and stability. To address this question, a series of experiments were conducted to test how mycelium reacts to various textures. These experiments aimed to determine which textures provide the mycelium with an appropriate surface to latch onto, while also maintaining adequate aeration, temperature and humidity necessary for its growth. A series of small-scale experiments were initially conducted to test the organism’s compatibility with the fabricated skin’s width distance, internal void size, geometry, skeleton and texture. The goal was to determine the ideal form for the mycelium to latch onto and provide sufficient oxygen and nutritious substrates to grow. Once the mycelium was grown over the fabricated form, a visual inspection was then applied to determine which of the coarseness–density measures was more successful in supporting the mycelium’s latching. Pleurotus ostreatus was utilized as it not only grows at a rapid pace but produces fascinating patterns of growth when searching for nutrients. From the set experiments, it seemed that mycelium was most attracted to intricate textured geometries rather than smooth surfaces. Scientific research has demonstrated that hyphae (the mycelium’s forming structure) can sense the texture of surfaces and detect grooves half a micro-metre deep. Furthermore, if hyphae are confronted with an obstacle, they will divert themselves around it by branching out and then back together (Sheldrake, Reference Sheldrake2020). By studying the organism’s behavioural patterns of growth, this research posits a strategy in which empirical observational knowledge becomes an input into the form-making process. Mycelium’s growth aesthetics is therefore converted into a language for architects to design with, catalysing a synergy with the biological. As a result, mycelium growth is utilized as a design generator and also for the fabrication of living, breathing architectural forms (Figure 1).

Figure 1.

Prototype 01: Mycelium growing inside recycled PETG skins.

In order to computationally generate intricate and textured surfaces, an algorithm based on swarm intelligence was deployed. This algorithm was specifically utilized to array and organize hundreds of components across the wall’s surface. Swarm intelligence is based on the collective behaviour of self-organized systems (Chu et al., Reference Chu, Huang, Roddick, Pan, Jędrzejowicz, Nguyen and Hoang2011). This behaviour is evident in natural systems such as ant colonization, birds flocking and wasps building nests. Computationally, algorithms based on the logic of swarm intelligence seek to mimic the robustness, adaptivity and efficiency of these systems, which emerges as a result of multiple interactions between simple components (Buus, Reference Buus2006). A number of agents therefore act according to very simple rules, governing the construction of large architectures. In order to generate rippled, intricate surfaces, each component (which are referred to as scales) was modelled at the dimensions of 70 mm height, 50 mm width and 20 mm depth. Through a swarm intelligence algorithm, these scales arrayed across the wall’s surface to simultaneously give the allusion of rippled skins and encourage both biological growth and mycelium’s adhesion to the fabricated skins. The wall’s surface was designed in varying degrees of coarseness, to support the mycelium’s ability to latch onto the skins. High-resolution intricate scales were arrayed across the fabricated skins at three contrasting levels of densities: low – scales at 50 mm apart; medium – scales at 20 mm apart; and high density – scales next to each other. Figure 2 presents these scale-like structures at the three density levels. Through observations, the growth and the reaction of the mycelium to the fabricated skins were monitored.

Figure 2.

An algorithm based on swarm intelligence was developed to array components across the surface, generating a reptile scale-like texture.

To enhance the tensile strength of the wall panel system, a methodology was developed by which external passageways were computationally designed to house fibre composites and provide additional tensile strength. As demonstrated in Figure 3, Fibre hemp ropes were laced within grooves prefabricated within the scaffold. These natural fibre ropes are 6 mm in diameter and with a breaking load of 285 kg, thus appropriate for adding the required tensile support. These grooves were then sealed and coated with a bio-resin, which is a plant-based non-toxic resin.

Figure 3.

Fibre hemp ropes laced within grooves prefabricated within the scaffold for additional tensile strength.

With the rise of machines and access to technology in our vicinity, robotic fabrication has become a widely accessible tool within the design and construction industry. Evident in ‘SensiLab’ and RMIT’s Building 515 project, Roland Snooks has proven that 3-D printing has now shifted from the small-scale prototyping stage into the commercialization of viable building products (Snooks, Reference Snooks2021). With this technology now meeting building codes and fire regulations, architects are retrofitting existing buildings with 3-D printed facades and wall panel systems to minimize material waste and decrease human labour. The downside however to utilizing this innovative technology is the common use of petroleum-based materials such as polyethylene or polypropylene (Lee et al., Reference Lee, Mohd Pu’ad, Selimin, Manap, Abdullah and Idris2020). With the motivation to reduce and reuse synthetic materials such as PETG, this research re-uses and re-purposes existing building materials. Designed by Roland Snooks, ‘Floe’ was comprised of seventy robotically fabricated PETG panels and presented at the National Gallery of Victoria (NGV) in 2018 (Snooks, Reference Snooks2021). In order to reuse and repurpose this plastic, each panel was ground into the form of pellets, which were then melted and extruded through a large-scale industrial robot – the Kuka KR150 robot. This same plastic was then utilized to 3-D print BioForms, a two-metre high wall panel system.

The final generated form was 3-D printed using recycled PETG. As demonstrated in Figure 4, a large-scale industrial robot, the Kuka KR 150, was used for 3-D printing. The form’s dimensions were 2000 mm in height, 800 mm in length and 30 mm in width. The skin thickness was 2 mm, and the internal gap was 80 mm. The fabricated skins were then painted with gradients of blue to visually enhance the form’s rippled, scaled and textured surfaces.

Figure 4.

The final fabricated wall panel system, 3-D printed from the Kuka KR150 robot.

Once 3-D printed, the robotically fabricated skins were ready for the application of the mycelium. The mycelium was then mixed with upcycled agricultural waste comprised of mycelium 60%, coffee 5%, paper pulp 25%, hemp seeds 5% and flax seeds 5%. This mixture was found to be effective in providing nutrients for the fungus’s growth. The scaffold was incubated for 14 days in a specially built greenhouse, keeping the humidity at 85% and the temperature at 30 °C, with natural light and airflow. The final form was left uncapped to encourage further growth of the mycelium. As the mycelium was kept alive, it continued to grow and fruit beyond the scaffold’s framework, forming long cylindrical tubes of oyster mushrooms. As demonstrated in Figures 5 and 6, the external fruiting of the fungus flourished and grew over time. During this process, mycelia began to form tiny networks of living labyrinths, filling all the available space between the fabricated skins. After a 14-day period, the fungi and nutrients formed a solid structure, as the organism adopted the shape and adhered to the two fabricated skins. Aesthetically, the organism embodied a thick cumbersome texture, mimicking the texture of foam.

Figure 5.

Mycelium growing inside the wall panel system.

Figure 6.

Mycelium growing inside the two-metre high wall panel system.

Exhibited as part of the 2021 NGV’s ‘Melbourne Design Week’, BioForms I was presented as part of the Printed Tectonics exhibition, displayed at the RMIT design archives window. The exhibition explored the positive environmental impact of building-scale 3-D printing and its implications for architectural design.

BioForms II technical workflow

Reflecting upon the BioForms I designed form, a few insights were revealed, which informed the development of BioForms II. First, it appeared that the plastic-based PETG scaffold was too smooth in texture for allowing successful latching of the mycelium. It was only in the high density of scale-coarseness that the mycelium was able to latch onto the skin. Thus, the final outcome did not appear as a unified form, but rather as loosely integrated, where the internal mycelium core seemed unstable as it was unable to successfully grip onto the surrounding fabricated skins. The findings also reveal that aesthetically, structures at a macro-scale of metres inhibit the expression of the fungus’s patterns of growth.

To address this problem, there was a need to turn to other materials, leading to the experimentation of wood-based skin solutions. When researching ideal materials to host mycelium growth, the earth’s natural nutrient cycles and processes of upscaling organic waste were explored. According to existing scientific research, fungi naturally consume nutrients contained in lignocellulosic biomass, which is the dry plant matter. These are often found in agricultural waste and fibrous substances such as straw, paper pulp, tree bark and sawdust (Kavanagh, Reference Kavanagh2005). To digest these materials, the fungi release enzymes from their hyphal tips, which result in intricate tapestries of growth (Jones et al., Reference Jones, Mautner, Luenco, Bismarck and John2020). Converting this scientific research into the architectural field, wood–plastic composites (WPC) were utilized in order to provide nutrients for the organism to consume and eventually degrade. Wood-based scaffolds seemed appropriate, as this material is used as a natural substrate for fungus, and it has a more penetrable surface. Reflecting on the first research stream BioForms I, from an aesthetic perspective, by enclosing the mycelium between the skins, it adopted a thick cumbersome foam-like appearance. Its intricate labyrinth patterns therefore became lost in the sheer volume of the material. To address this problem, there was a need to allow the mycelium to grow externally and contribute to the aesthetics of the overall formation. In summary, three main changes were made when progressing from BioForms I to II. These were changing in skin material, creating scaled textures at high density across the entire surfaces of the skin and allowing the mycelium to form its patterns of growth external to the wall panel. In BioForms II, the following changes were implemented in the creation of the form: (i) The scaffold was made of wood–plastic composites (WPC), composed of corn starch (PLA [60%]), and wood cellulose (40%); (ii) the scaled density was now set to be high and uniform across the skins; (iii) mycelium culture was infused on the external skin and not only applied to the interior of the skin; and (iii) the shape of the structure advanced past the standardized wall panel system into baroque rationality through high-resolution organic forms that further exposed my architectural design aesthetics. These changes were derived as outcomes from reflecting on the design outcome of BioForms I.

In order to enhance the organism’s adhesion to the designed surface, skin textures were amplified, providing coarseness for the organism to latch onto. Rather than arraying scales across the external surface wall, a new algorithm was generated which morphed the surface both internally and externally. Utilizing an algorithm based on cellular growth, complex organic structures were generated to create intricate groves and textured pathways for the organism to seep through and spread its hyphae tips. As demonstrated in Figure 7, through these heterogeneous skins, form varied in areas of intensity, sparsity and scales, generating pathways for the organism to seep through and expose its agency of growth. The developed algorithm utilized the principles of differential growth, developed by Alan Turing in his paper ‘The Chemical Basis of Morphogenesis’. Here, a computational cell receives nutrients from its environment, which alters its overall network topology (Turing, Reference Turing1990). Structures were created from interconnective cells with rules for the forces between the cells and for how the cells accumulate internal nutrients. When the nutrient level in the cell is given a threshold, the cell splits into two and reconnects with its immediate neighbour (Lomas, Reference Lomas2014). Results show an incredible similarity to biological organisms, which were specifically tailored to embody a sharp contrast between smooth and intricate areas. As a result, a catalogue of intricate, textured surfaces were computationally designed to simultaneously encourage mycelium growth and expose my architectural aesthetic of baroque formations and amplified resolution. These intricate forms not only operate performatively as they enable mycelium to grow and adhere in certain ways but also expose the architects’ design aesthetics of high resolution and delicate complexity. These heterogeneous skins transitioned from smooth into highly detailed surfaces, embodying organic forms that are ordinarily seen in nature. These skins were intentionally restricted to a 40-degree angle to ensure successful printability on the robot.

Figure 7.

A catalogue of intricate, textured surfaces computationally designed to simultaneously encourage mycelium growth and expose my architectural aesthetic, utilizing an algorithm based on cellular growth.

To test the ideal scale, a series of wall panels were designed with contrasting gap sizes between surfaces. The question posed was: What gap size between the two fabricated skins is appropriate for obtaining maximum structural support for the form while maintaining the mycelium’s properties of growth? The appropriate gap was determined as such when it ensures the adhesion of the mycelium and its compact enclosure between the skins, supports the mycelium’s weight and allows oxygen to permeate the model, so as to not suffocate the mycelium. In addressing this question, I experimented with various gap sizes. The three trailed gap sizes were 50 mm, 40 mm and 20 mm. Sections were extracted from each computational model and fabricated on a small-scale 3-D printer in order to test these qualities on an accurate scale. The skins were 3-D printed using a Prusa 3-D printer, with a printing bed size of 300 mm × 300 mm × 30 mm. This printer is suitable for small-scale projects, as in my three trial experiments. The WPC type used in this printer is polylactic acid or polylactide (PLA) derived from fermented starch (Lamm et al., Reference Lamm, Wang, Kishore, Tekinalp, Kunc, Wang, Gardner and Ozcan2020). The filament was extruded through a 0.8 mm nozzle to prevent the wooden fibres from clogging the extruder.

Once the form was fabricated, I used a liquefied culture of mycelium for infusing on the external skins. The form was then incubated for 14 days, as the mycelium grew within the internal gap as well as on the external textured surfaces. Figure 8 demonstrates mycelium’s growth in the enclosed gap between the two fabricated skins and its exposure on the skins’ high-resolution textured surface. Internally, the organism embodies a thick cumbersome texture, while externally the organism exposes its intricate agency of wispy fibrous growth. It was therefore found that the dual applications of using WPC to create the scaffolds and applying scales-texture at high density across the entire surfaces were effective in improving the latching of the mycelium onto the skins. Aesthetically, the organism exposed its wild patterns of growth as it flourished through the uncapped voids in each digitally fabricated model. The findings revealed that the appropriate width of the gap is 40 mm, met the requirements of an adhesion of the mycelium and its compact enclosure between the skins, supported the mycelium’s weight and allowed oxygen to permeate the model, as to not suffocate the mycelium.

Figure 8.

Mycelium growing inside the 3-D printed wooden wall panel system.

Building on these results, these initial tests were then scaled up to create the final form to the scale of metres. The final computational form was designed as a team, incorporating these findings of high-resolution skins and internal gap size to encourage mycelium growth. Here, the geometry embodies textured grooves for the organism to latch onto, and internal pathways to seep through, exhibiting coral-like (Cnidaria) formal qualities, often seen in marine environments. Here once again, the Kuka KR 150 robot was used for 3-D printing the skins. To do so, WPC pallets were dried at 100 °C for 4 hours to reduce the moisture content to 2%. This ensured that no air bubbles were formed during the extruding of the material. Once inserted into the printer, the material was further heated at 170 °C for 3-D printing. Arbitrary geometries were initially generated in order to test the fibrous material’s flow rates, melting temperature, layer heights and compatibility with the Kuka KR 150’s extrude.

Similar to the inoculation process of BioForms I, mycelium was grown within the internal walls of BioForms II, acting as a natural binder to adhere the fabricated skins. Thus, BioForms II was successful in demonstrating the formation of a unified wall panel, where the mycelium acted as the internal natural glue, bonding the fabricated skins together to form a sandwich-style panel system. However, being kept alive, the fungus often strayed away from the pre-designated enclosures, affirming its own sense of agency and volatility. Unlike the growth within the internal surface of the skins, which was homogenous, appearing as a thick cumbersome foam, the examination of the external walls revealed entirely different aesthetics. Here a rich heterogeneous silhouette of formations were apparent, presenting a catalogue of geometric features, such as high-resolution textured surfaces, mesh mutations and internal folds. As demonstrated in Figure 9, the form embodies features of organic growth, aesthetically blurring the architectural intention by infusing the fungus’s autonomous agency into the pre-determined design of the wall panels. It was therefore found that the dual applications of using WPC to create the scaffolds and applying scales-texture at high density across the entire surfaces were effective in improving the latching of the mycelium onto the skins.

Figure 9.

Mycelium growing throughout the 3-D printed wooden skins exposing contrasting biological patterns of growth.

While the fabricated solid surfaces were 2 mm thick, the organism at times burst through these solid surfaces and began to fruit, exhibiting a wispy delicate texture. Finally, there was a need to address the challenge of scaling up and controlling the mycelium’s growth. Here, the following question was posed: To what extent can the mycelium’s growth patterns be scaled up and controlled through fabrication, while keeping the mycelium alive and thriving? In relation to scale, it was found that once scaled up to metres, aesthetically the mycelium takes on a thick, cumbersome form, masking its intricate details that were once visible at the micro-scale. Similarly, the living organism maintains its volatility, regardless of attempts to control it. The experiments revealed that regardless of applying rigid computational guidelines and imposing digitally fabricated enclosures, the fungus could not be completely controlled. While there were some sections in the living prototype that were successful in controlling the mycelium, in other parts, it strayed away from these rigid guidelines, asserting its own agency. Figure 10 demonstrates the fungus’s self-assertion as it deviates from the set frame.

Figure 10.

Mycelium bursting through the 3-D printing framework, asserting its own agency within the generated form.