Page 1






FACULTY “More than a substance, plastic is the very idea of its infinite transformation; as its everyday name indicates, it is ubiquity made visible. And it is this, in fact, which makes it a miraculous substance: a miracle is always a sudden transformation of nature. Plastic remains impregnated throughout with this wonder: it is less a thing than a trace of movement. – Roland Barthes

Areti Markopoulou


ASSISTANT Angelos Chronis

RESEARCH TEAM Noor El-Gewely Lili Tayefi Christopher Wong


01 12 02 03 04 05 06 07 08 09



16 60 72 92 104 122 124 128

Programming Material Intelligence Using Food Waste Deposition to Trigger Automatic Three-Dimensional Formation Response in Bioplastics



While contemporary plastic manufacturing is readily available and relatively cheap, it does not attend to environmental issues. With the growing global demand for energy, chemicals and materials, bioplastics offer several advantages including sustainable production and consumption, using renewable raw materials to meet global targets. Food waste contributes to excess consumption of freshwater and fossil fuels which, along with methane and carbon dioxide emissions from decomposing food, impacts global climate change. With the goal of a sustainable bioplastic, this paper explores a programmable food waste composite, thereby using waste relative to local production. The fabrication methods used include embedding self-assembly properties within the material program to create semi-autonomous plastic panels as a result of dehydration. Such panels can be used in various scales from product design to large scale, temporary, local ephemeral architectural applications. This study focuses on the exploration of the programming of material intelligence via digital and physical simulations, using food waste deposition to trigger a self-assembly response in bioplastic. Previous work on biodegradable plastics has been researching different mixes for enhancing mechanical properties (Kretzer, et al. 2013) but as well as variation of mechanical properties in anisotropic surfaces based on multi-material deposition. The current research additionally focuses on self-assembly processes based on the material properties of the new composite.

Exploration of the potential of bioplastic as a computationally designable and digitally fabricable material was conducted along two distinct but interrelated axes: investigation of the components comprising the bioplastic and their proportions – the recipe – was performed concurrently with the design of processes and equipment for fabrication using the material. A recipe and a process which, when applied together, form a system with a unique performance and the potential for application in the fabrication of low-cost, large-scale complex geometric architectural surfaces. This system enables the automatic formation of a rigid bioplastic three-dimensional hyperbolic paraboloid from a two-dimensional sheet cast. Materially, the system combines a gelatin-based bioplastic with granular organic matter from food waste, which confers specific characteristics depending of on the type used. This material is cast into a flat frame laser-cut from medium-density fiberboard, which is split at two points to enable the bioplastic’s self-formation into a three-dimensional doubly curved panel. Experimentation was performed with the aims of achieving control over the degree of deformation of the material and, in parallel, empirically quantifying this action in relation to its informing variables in order to develop a parametric model and simulation with which to investigate formal possibilities and potential architectural applications.


Bioplastics are by their very nature parametric materials, programmable through the selection of constituent components and the ratios in which they appear, and as such present significant potential as architectural building materials for reasons beyond sustainability and biodegradability. This paper presents a system through which rigid three-dimensional doubly curved hyperbolic paraboloid shapes are automatically formed from two-dimensional sheet casts by harnessing the inherent flexibility and expressiveness of bioplastics. The system uses a gelatin-based bioplastic supplemented with granular organic matter from food waste in conjunction with a split-frame casting system that enables the self-formation of three-dimensional geometries by directing the force of the bioplastic’s uniform contraction as it dries. By adjusting the food waste


added to the bioplastic, its properties can be tuned according to formal and performative needs; here, dehydrated granulated orange peel and dehydrated spent espresso-ground coffee are used both to impart their inherent characteristics and also to influence the degree of curvature of the resulting bioplastic surfaces. Multi-material casts incorporating both orange peel bioplastic and coffee grounds bioplastic are shown to exert a greater influence over the degree of curvature than either bioplastic alone, and skeletonized panels are shown to exhibit the same behavior as their solid counterparts. Potential developments of the technology so as to gain greater control of the curvature performance, particularly in the direction of computer-controlled additive manufacturing, are considered, as is the potential of application in architectural scale.



2.1 BASE BIOPLASTIC COMPOSITION Extensive testing of bioplastic recipes was performed, with particular attention paid to interactions with potential food waste particles and fabricability through a variety of methods including two-dimensional casting, three-dimensional casting and syringe extrusion, all at a variety of scales. (Figure 1 & 2). As the tests of both material and technique iteratively converged on the system described above, the following base bioplastic recipe – adapted from Materiability (Rodriguez 2012) with the addition of vinegar – was found to offer the most synergistic pre-casting, post-casting and post-curing properties among those tested:  20 parts (by mass) water  4 parts dry gelatin powder  1 part glycerin  1 part vinegar To prepare the bioplastic, these ingredients are added to a pot in the order listed and heated on a hot plate while being stirred continuously until the gelatin becomes amorphous (between 40 and 60 ºC). The process of deriving this recipe emphasized the parametric nature of plastics: each constituent played an observable role in the outcome of the material. For this recipe, gelatin and glycerin form


the basis of the plastic, lending among other qualities transparency and cohesion respectively, while vinegar rigidifies the cured result. Previous iterations using ingredients such as corn starch and silicone in large proportions were also revealing of plastics’ ability to take on their components’ characteristics, though their effects were not always desirable. For example, starch-based recipes remained glutinous despite weeks of curing time and attracted populations of Drosophila, while recipes containing silicone were not evenly miscible and readily became moldy. Overall, the gelatin-based bioplastic conferred the best mix of characteristics among the recipes tested, and was chosen for its stability, transparency, post-curing rigidity, and compatibility with the split-frame fabrication system.

2.2 ADDITION OF GRANULAR ORGANIC MATTER FROM FOOD WASTE The addition of organic particles to the bioplastic was initially investigated as an avenue by which to augment its characteristics through incorporating those of other materials. Granulated orange peel, for instance, was trialed as a means to integrate the structural properties of cellulose – namely strength (Genet, et al. 2005), rigidity (Bayer, et al. 1998) and bulk – through the use of an abundant and largely unused food waste. Simultaneously,

spent espresso-ground coffee was tested with the aim of enhancing the water resistance of the bioplastic via unextracted flavor compounds, which are hydrophobic (Wouda 1981). Tests of bioplastic samples impregnated with granulated orange peel and spent coffee grounds – hereinafter referred to as orange peel bioplastic and coffee grounds bioplastic respectively, for brevity – were performed in parallel on samples containing increasing amounts of dehydrated food waste (five, 10 and 20 percent by mass), alongside a control containing no food waste. Each test was performed on a sample created by casting into a 100 mm square, 3 mm deep MDF mold. Among the tests performed, two in particular revealed generalizable results:  Subjecting the samples to one minute of heating from a 600 ºC heat gun from a distance of 100 mm exhibited deformation inversely proportional to orange peel content; and  Casting in a diagonally split frame exhibited deformation proportional to food waste content upon drying, for both orange peel bioplastic and coffee grounds bioplastic – though the latter showed more acute deformation. In this way, the added granular organic matter from food waste became a major parameter in its own right, with both the type of waste used and the amount used influencing the characteristics of the bioplastic.

Additional testing revealed structural potentials in the orange peel bioplastic. Three equilateral triangular sheets with a side length of 150 mm, cast at 5 mm thick and containing food waste particles, were suspended from their vertices and allowed to dry into catenary arches – as in Gaudi’s technique (Huerta 2006) – and inverted. Load was incrementally applied using weights, with the arch containing containing 30 percent orange peel successfully bearing 18.45 kg without collapsing, and rebounding to its original shape within 24 hours. 2.3 SPLIT FRAME DESIGN AND AUTOMATIC FORMATION The automatic formation of a doubly curved panel in this system is a result of the bioplastic and the frame in which it is cast working in concert. As the bioplastic dries, the egress of water through evaporation causes the material to contract. Left unconstrained, this contraction occurs anisotropically and to a degree proportional to the bioplastic’s food waste content. However, if the bioplastic is cast into a frame such that its edge conditions are regulated, then it is possible to passively exert control over the result of the contraction. By casting the bioplastic into split frames laser-cut from medium-density fiberboard, an axis of freedom which dominates the bioplastic’s natural anisotropy can be established.



This effect is a function of both the frame’s material and its geometry. Because the bioplastic adheres to the porous fiberboard, its edge condition is regulated where it contacts the frame, thereby constraining the material’s contractive tendency. Any points of weakness in the frame therefore become outlets for the force of contraction – thus, by splitting the frame at two points, an axis of freedom is inscribed between them, around which the material is free to bend as it contracts. Experimentation has revealed a number of points to consider pertaining to this technique. Firstly, the bioplastic should be cast upon a minimally porous surface for ease of release. Testing has shown that a cast bioplastic surface, once sufficiently solidified as to be non-liquid – typically after one hour of drying – can be peeled from the casting surface and hung up vertically to cure with minimal effect from gravity on the geometry of the result. Secondly, the curvature of the edges of the panel is influenced by the rigidity of the frame; for a given panel size and to a limit, a frame cut from thicker fiberboard yields straighter edges.


2.4 INVESTIGATING THE NATURE OF CONTRACTION With a technique enabling the formation of three-dimensional hyperbolic paraboloid surfaces from two-dimensional panel casts established, experimentation was performed to explore methods by which the degree of bending might be controlled. Early split frame tests indicated that, for a given proportion of food waste matter, panels cast from coffee grounds bioplastic contracted more than those cast from orange peel bioplastic. To investigate whether the materials’ different contraction rates could be leveraged to influence the curvature of a panel, contrasting multi-material casts were made: one featuring a strip of coffee grounds bioplastic running from one corner split to the other with orange peel bioplastic comprising the rest; and another with the same materials in the opposite orientation. The results of the test were dramatic: at a 100 mm square panel size, the coffee grounds panel with the orange peel strip contracted to roughly a 100 degree angle when viewed from the side, while the opposite panel contracted to only approximately 30 degrees. This contrast is considerably more marked than that between mono-material panels, indicating that the two bioplastics can be used in conjunction to either amplify or temper the automatic formation effect.

In addition, tests were performed to assess the geometry of the contraction itself, with a view to parameterizing the deformation. In the next series, the grid was skeletonized – that is, a mold was created so that the bioplastic was cast only on the grid lines, at a width of 7.5 mm, with the cells of the grid left empty – to observe whether the introduction of void space affects the pattern of contraction. Both series of tests showed that both bioplastics contract uniformly on a per-material basis, with neither the edges nor the center of the panels contracting materially faster or more than the other.




Currently the main industrial application of bioplastics is in disposable objects, such as packaging, cutlery, or shopping bags. There are a few precendent projects which use bioplastics for architectural applications. In our project we wish to investigate the implications of using bioplastic at an architectural scale. And how the temporary nature of bioplastic might respond to its environment and change over time. Another aspect in our research is how we fabricate our designs using additive manufactering at a large scale.


Gernot Rainer AA Pavilion


PRECEDENT PROJECTS DUS Architects used a bioplastic facade for the Mobile Europe Building in Amsterdam. The bio-plastic can be fully recycled after the six-month presidency is over. This particular bio-plastic was made from linseed oil. The facade was printed locally using a XXL 3D, that can print 2 x 2 x 3.5m. The prototype was created using fused deposition modelling, which is an additive manufactering technique employed by most domestic scale 3D printers. The AIA Pavilion in New Orleans by Gernot Reither is an exmple of how plastics can be used in an affordable and environmentally friendly way. By working in collaboration with scientists in chemistry, architects can work at the scale of molecules to design to redistribute the strucure of matter. Plastics can be considered a parametric material, since its properties can be programmed to suit based on the desired characteristics. The AIA pavillion was made from a glycol-modified polyethylene terephthalate (PETG), which can be produced from recycled plastics or sugar cane. The sugar cane plant has been an itegral part of the Louisiana culture and economy for 200 years. There are environmental benefits in producing PETG from sugar cane. Braskem, a Brazilian chemical group claims that producing one ton of sugar cane ethanol to make polyethylene, removes 2.5 tons of carbon dioxide from the atmosphere, whilst the traditional petrochemical production of plastics would contribute to 3.5 tons of emissions. The ICD ITKE Research Pavilion 2013-2014 was a reference for us in terms of a large scale robotically fabricated prototype, using fiber reinforced polymer. This is an example of how through differentiation of the material deposistion one can achieve a lightweight structure using minimal material.


Europe Building - DUS Architects

AIA Pavillion 2011 - Gernot Reither

ICD ITKE Research Pavilion 2014

KamerMaker - DUS Architects

AIA Pavillion 2011 - Gernot Reither

ICD ITKE Research Pavilion 2014


The research studio Digital Matter Intelligent Constructions [DMIC] began with a week long workshop called: Space Dynamics, which was led by Manuel Kretzer. It was during this workshop that we were first introduced to the concept of Bioplastics, among other smart materials. The aim of the workshop was to familiarize ourselves with new materials, as well as a hands-on methodology of experimenting and prototyping. We were given some base recipes to test of various kind of bioplastics, which are also published on the Materiability website. There were two strategies of approach we could take with material exploration; the first to select a material performance and select a material to achieve this, the second option was to experiment with a material, then discover its behaviour and potential. For our project we sought the second option.


In this chapter about Material Exploration we will cover our initial experiments with bioplastic. Early on we tested various recipes and combinations of ingredients in search of a base material to work with. Our next step was to incorporate a granular food waste into the bioplastic. Then we proceeded to subject the material to a series of stress and stain tests - such as heat, water, percentage of food waste. From these experiments the main objective was to observe the materials transformation over time. Bioplastic being a biodegrable material is sensitive to fluxtuations in temperature and humidity. One of the main differences between bioplastic and traditional plastic is that bioplastic is dynamic, and is subject to transformations over time. Once we are familiar with the resultant behaviours we can begin to design a material system.





PARAMETRIC MATERIAL et et as quam, voluptios poreperum rerat quassum que nihilit aepedit iatur, ulparum accus. AFFORDABLE To doloribuscim non parita quost eumReris conseque imi, corrovid quid quid unt, to ipis aut ACCESSIBLE et elest quae eatia nobis maxima quiata quiam, sus audipid que cum con eostium quiam fac cuscient alicatur autas eicienditiis doluptat pero omnis doloris cipiendi dio di aciam entoLIGHTWEIGHT eum nulliaectio. Uptatum aut eati nulpa conseque de nat. NON-TOXIC Debitas repro qui desequam enis aut es sitin rehenientem et mo mo derum imoluptae porite RIGID FLEXIBLE labo. Ovid quam nis sim repe pelestisit fugit offic tem que rerspic iasped qui dolorate num TRANSPERANT quamus corerum is qui as millendam, alia de nat. Sequistrum nobis iur? Ecatet fugit ipsanis aut que volessita doluptatur, quae explis et,HEAT suntur RESPONSIVE sinvel idunt harchicidunt volorum erferib uscimusam quiandio. Lor re, quatentemporDESIGNER/MAKER aut aut quati tota essitia nihil ium ariosse nderspe lluptat umquam, et maximax imincim quia nos CAN CONTROL am quos et qui adias utectat ullati verae laccuptam fugit, odit et parum rempos et aborernam

parciendus essumendi conserf erumque qui tem etur aut quidebissi





[1] corn starch 18.01.2016 11:36 2 parts water 2 part cornstarch 2 parts glycerine 1 part vinegar Flexible, bendable Dries well Dries in printed format

[2] gelatine 19.01.2016 14:00 240 mL water 48 g gelatine 12 g glycerine Hard Transparent Very heat sensitive Smooth

[3] corn starch + silicone 19.01.2016 14:30 100 g corn starch recipe + 91 g (A) translucenet silicone mold rubber 9 g (B) silicone rubber compound

[4] gelatine + silicone 20.01.2016 11:36 100 g gelatine recipe + 91 g (A) translucenet silicone mold rubber 9 g (B) silicone rubber compound Still flexible; not dry after 1wk Mold after 1 week Printed grainy due to reheat

[5] gelatine + cornstarch + silicone 20.01.2016 11:36 30 g gelatine recipe 30 g corn starch recipe 30 g silicone Shrinks in the short direction Composite properties High rigidity/stiffness/ strength

BASE RECIPE In this series of experiments we were testing various recipes to determine their characteristics both in cast and extruded format. At this stage we were trying to identify the base recipe of our bioplastic. The gelatin-based recipe was selected for its rigity and smooth surface finish, as well as transperancy. The extruded test was unsuccessful, however this was due to the temperature being too high at the time of printing.



MATERIAL EXPLORATION In a presentation by Ilker Bayer we were introduced to the concept of using food waste as one of the components of the bioplastic ingredients. Depending on which food waste is used, different properties can be harnessed and incorporated into the bioplastic. Food waste is an abundant material, which is annually renewed, and it is essentially free or low cost. Food waste is currently created at a large volume, and is generally perceived as a unwanted material. At both an agricultural stage and at processing and consumption stage there is an oppurtunity to utilize this material. This has the added benefit of diverting this matter from landfills and the transportation and logistical requirement needed to do so. In our selection of food waste, we initally tested types of food waste that were geographically specific to Spain, and also particularly Barcelona. Spain is famous for its oranges, and Barcelona being a coastal town has plenty of seafood. So, therefore we selected orange peel and shrimp shells as additives. The idea behind using a locally sourced material, is that the end user would be able to use any food waste matter that is available to them. By adding a food waste to the bioplastic it increases the bulk of the material, and also reduces the cost. Orange peel was also selected for being a good source of cellulose, which offers strength and heat resistance properties than the gelatin-based bioplastic would have on its own. The food waste is collected, cleaned and then pulverized and added into the base recipe. Both the orange peel and the shrimp shell were tested with two types of bioplastic - gelatin and silicon based. At the time of casting both seemed to produce simialr results. However one week later, the samples had aged very differently. The silicone bioplastic sample had started to mold, and discoloration. From this we excluded silicone from our sample set of future tests due to it bio-incompatibility with organic matter. Another important lesson learned was that biological materials tend to transform over time.





[7] gelatin + corn starch + orange peel

[8] geltin + silicone + orange peel

[9] gelatin + corn starch + shrimp shells

[10] geltin + silicone + shrimp shells

[8] geltin + silicone + orange peel

[9] gelatin + corn starch + shrimp shells

[10] geltin + silicone + shrimp shells

[time] 1 time 0d 28.01.16

SHRIMP SHELLS Shrimp shells were used for chitin, which is tough yet flexible material. It is found in many crustaceans and insects with exoskeletons. However this sample was extremely unpopular with our fellow resear-

[7] gelatin + corn starch + orange peel

chers, due to the strong odor . Perhaps the shells were not adeqautely cleansed, and therefore began to rot. The orange peel definatly produced a more plesant aroma.


[time] 1 time 11d 08.02.16






In this experiment we tested 3 types of orange peel bioplastic, combined with corn starch, gelatin and silicone. These samples were then hung to dry forming catenary arches. The gelatin based recipe had the most uniform surface finish, and material consistency. Also it was desirable due to its transperant quality.




STRENGHT TEST In the previous experiment the material sample had cured into a catenary arch shape. In this experiment we wanted to test the strength of this arch, since the material seemed to be quite rigid yet flexible. The sample size tested was a triangle with edge length of approximately 150mm length. We progressively added more and more weight to test the fatigue of the material. As the load increased, there was a deformation to the arch, it flattened out. However although the deformation appeared to be global, there were no fractures in the surface. And also the shape rebounded back to its original shape after 24 hours. This implies a material that relatively strong compared to its size and also flexible.












Pre-heat testing

In this experiment we subjected the bioplastic to heat from a heat gun of up to 600deg Celsius at a distance of 30cm, for 3minutes. The bioplastic surface began to melt, and eventually burn with the excess material of granular orange peel fall off. The test sample of bioplastic without OP began to burn and bubble in a shorter amount of time. The consulting material scientist, Ilker Bayer said that the food waste would infuse it’s properties to the base recipe. In this instance the addition of orange peel improved the heat resistance capabilities. This is due to the cellulose, which does not become aqueous untill 370degrees.


Post-heat testing



[parameter] none original / 0 / 5 / 10 /20%

[parameter] frame split curing time 48hrs original / 0 / 5 / 10 /20%

[parameter] splits curing time 24hrs original / 0 / 5 / 10 /20%

[parameter] heat 1min applied at 100mm original / 0 / 5 / 10 /20%

[parameter] water soaking for 9hrs original / 0 / 5 / 10 /20%




[parameter] none original / 0 / 5 / 10 /20%

[parameter] frame split curing time 48hrs original / 0 / 5 / 10 /20%

[parameter] splits curing time 24hrs original / 0 / 5 / 10 /20%

[parameter] heat 1min applied at 100mm original / 0 / 5 / 10 /20%

[parameter] water soaking for 9hrs original / 0 / 5 / 10 /20%




65% @ 65.00% T: + 1wk 80.00% 80% @ T: 48hrs

100% @100.00% TIME OF CAST In our material explorations we have observed that the input of hydation and the output of dehydration affects the size, shape and mechanical properties of the food waste bioplatic. The material is cast in a liquid state, and as it dries it tends to shrink and curl at the edges. If the material is exposed to water after it has been cured it can expand and soften again.




am quos et qui adias utectat ullati verae laccuptam fugit, odit et parum rempos et aborernam et et as quam, voluptios poreperum rerat quassum que nihilit aepedit iatur, ulparum accus. To doloribuscim non parita quost eumReris conseque imi, corrovid quid quid unt, to ipis aut et elest quae eatia nobis maxima quiata quiam, sus audipid que cum con eostium quiam fac cuscient alicatur autas eicienditiis doluptat pero omnis doloris cipiendi dio di aciam ento eum nulliaectio. Uptatum aut eati nulpa conseque de nat. Debitas repro qui desequam enis aut es sitin rehenientem et mo mo derum imoluptae porite labo. Ovid quam nis sim repe pelestisit fugit offic tem que rerspic iasped qui dolorate num quamus corerum is qui as millendam, alia de nat. Sequistrum nobis iur? Ecatet fugit ipsanis aut que volessita doluptatur, quae explis et, suntur sinvel idunt harchicidunt volorum erferib uscimusam quiandio. Lor re, quatentempor aut aut quati tota essitia nihil ium ariosse nderspe lluptat umquam, et maximax imincim quia nos parciendus essumendi conserf erumque qui tem etur aut quidebissi


The concept behind this prototype was to design a system which works in parallel with the material performance. In this case this is a scale model of the responsive facade system made of bioplastic. When both the facade itself and the surrounding climatic condition are dry, then the facade component will be dry and curled allowing the apertures to be open. This will allow views and ventilation through.

However when there is rain, then the facade will get saturated and the facade components will expand and become heavy with water. And this will actuate a closed state of the facade. This will only work with certain temperatures of water (below 35 deg.), if it is higher then the bioplastic material can begin to disintigrate.




Various food waste additives were tested to compare their hydrophobic properties - orange peel, tomato skins and coffee grounds. Ten percent of food waste was added to each of these test. From the test we can see that the coffee is the most hydrophobic of these samples, due to the angle formed by a drop of water on the surface. Also coffee beans contain oils which help repel water. The tomato skin bioplastic also demostrated hydrophobic properties, however it required many tomatoes so it was quite labor intensive for our production method.

tomato peel bioplastic

Image Credits


coffee powder bioplastic




2. 3.



1. Orange Bioplastic Vein 10% 2. 3mm MDF Flex 1. jointOrange Bioplastic Vein 10% 3. Coffee Bioplastic Panel 15% 2. 3mm MDF flex joint 4. 3mm MDF casting frame 3. Coffee Bioplastic Vein 15% 5. 3mm MDF casting frame

4. 3mm MDF casting frame 5. 3mm MDF casting frame




We tested our Material System reversing both Orange and Coffee as the main panel material to observe the difference. We found that the Coffee bioplastic panel resulted in a greater degree of folding, and therefore a more dramatic transformation. Aestetically and behaviourally we found this quality appealing, a kind of self-forming panel. The reason for the shape-shifting is the loss of water content in the material as dehydration occurs. Due to Coffee’s hydrophobic properties, the effect is more prominent here.





B.. C.. 8.








3. 2.

Frame Suporting frame Sample base Manometer Pneumatic head Pneumatic piston Manometer Regulator Regulator Compressor conenction Sample p

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.. 11.

[parameters] stress, strain, young’s modulus

We utilized a machine developed by Sofoklis Giannakopoulos to test the tensile strength. And the following is his description of how it works: ‘Tensile testing, also known as tension testing, is a fundamental material science test in which a sample is subjected to a controlled tension until failure. The results from the test are commonly used to select a material for an application, for quality control, and to predict how a material will react under other types of forces. Ultimate tensile strength, often shortened to tensile strength (TS) or ultimate strength, is the maximum stress that a material can withstand while being streched or pulled before falling or breaking. Tensile strength is not the same as compressive strength and the values can be quite different. Some materials will break sharply, without plastic deformation, in what is called a brittle failure. Others, which are more ductile, including most metals, will experience some plastic deformation and possible necking before fracture. For that reason a crash machine had been fabricated. The crash machine uses a pneumatic piston connected in linear with two press regulators (each one connected with a manometer) and a compressor that can reach 11 bars. The pneumatic piston is a corsa 75 iso up to 10 bars with a diameter of 32mm. The first regulator (9) goes up to 11 bars and the second one (8) up to 4 bars with percission and a small bima. By applying to each of the casting samples a press till the fracture of the sample, it is possible to test and record the maximum flexibility and strength of each of the samples in Newtown/mm2.’


From the mechanical property tests performed by the IIT the test results ranged from 72MPa - 339MPa, which places the bioplastic in the range of other composite materials and similar to even some metals and alloys. One point to note is the mechanical properties are subject to change based on the humidity level.


Initial State: Hydrated Panel

48 Output State: Dehydrated Panel




Initial State: Hydrated Panel

t 8: 23:14:35 t 7: 20:25:13 t 6: 19:16:45 t 5: 15:58:33 t 4: 13:27:12 t 3: 08:17:17 t 2: 04:59:05 t 1: 00:57:39


Output State: Dehydrated Panel



From a design perspective there are different properties that the various types of bio-plastic offer us, based on what type of food-waste we incorporate into it. Due to the ease of production and accessibility of this material, we can tailor the desired parameters to suit, based on our requirements for each project, but also to fine-tune the material performance within any given structure or object. Since we are working with a malleable material we see a potential to utilize this material at various scales - from product, to installation, to architectural. Although we have chosen to experiment with coffee and orange thus far, we do anticipate that this bio-plastic can be locally re-appropriated to incorporate food-waste which is most abundant at a given geographical location.

[CP] - hydrophobic - shading [gel] - transperancy [OP] - strength



[OP] - strength



Potential Site Potential Site - El Born, Barcelona

Commercial Food Waste Collection > Hub

Residential Food Waste Collection > Hub

Bioplastic > Users

FOOD WASTE COLLECTION At this stage in the project we are investigating on how we can incorporate the material system of Piel Vivo into our urban cityscape, and participate in the food production and energy cycle of the city. We envision a strategy for the city of the future, where each neighbourhood could have a hub (Shown in Red) to collect food-waste from its surrounding, Food-sources (Yellow) and Residents (Blue). At the food-waste hub the matter is collected and sorted, and the bio-plastic is processed here. The bio-plastic can be made on-demand based on the user requirements, and go directly back for distribution to its local community, reducing the need for transportation.y for the city of the future, where each neighbourhood could have a hub (Shown in Red) to collect food-waste from its surrounding, Food-sources (Yellow) and TITEL Residents (Blue). the food-waste the matter is collected and Mo corae et quis At dolorro to dolorib hub usaperum re nonesto et quisEsorted,Ipiet, and the bio-plasticsecae is processed here. The bio-plastic can be pedio. unditatendia pelit asimus, offictius, quia exerume made on-demand based on the user requirements, andacilige go directly ea simillutem. Nequi corepera cuptatio. Igeniss inullut ndebis back for distribution itsvero localetcommunity, the need eatur, susdant quam,to qui es modiciis reducing nam, voluptate de for sit pel transportation. init omnihil ligent.







03 In this chapter we will focus on viscosity, as this is an essential property to get right before being able to move forward with extrusion based robotic fabrication. One of the very first test we needed to do was to check how the material will flow through the extruder with gravity alone. If it flows through, we need to adjust the material. The more viscous a material is the better for extrusion it is, because then it can be controlled by pressure. In order to adjust the viscosity of our material, we isolated each component of the recipe and varied them one by one, to determine which would influence the result. The factors we tested were - reducing water, increasing gelatin, increasing foodwaste, and adding starch. We found that increasing the food waste content had the most successful effect. Therefore we pro-



ceeded to test various percentages of food waste for both the orange peel and coffee bioplastic. The viscosity of the material was also proportional to the diameter of the nozzle size. Another important factor that we adressed at this stage of the project was the material uniformity. The less uniform a material is , the more difficult it is to print. In order to improve the material uniformity we needed to reduce the granular size of the orange peel. This was done by two steps: a better food processor, and dehydrating the OP in the oven before processing it. Concurrently with the viscocity tests, we were observing the effects of linear deposition rather than a solid surface. Would the material still perform the same behaviour when we changed how and where we would deposit it?













25% OP LINE 6MM dia.

50% OP LINE 6MM dia.

50% OP LINE 6MM dia.




25% CP LINE 6MM dia.

37.5% CP LINE 6MM dia.

31.25% CP LINE 6MM dia.

34.25% CP LINE 6MM dia.



04 Until this stage in the project we had mainly focused on a manual casting fabrication method. But one of our aims for this project, was to be able to robotically fabricate by the means of additive manufacturing. There are several parameters to control when it comes to additive manufactering such as - viscosity, nozzle diameter, temperature, air pressure, height above printing surface. In terms of our additive manufacteruring workflow approach, we began with manual techniques then moved towards automated: manual syringes > air pressure assisted extrusion > Kuka Sofoklis Small extruder > Kuka Sofoklis Big Extruder. This methodology allowed us to do quick prototy-



ping while we were learning about robotic fabrication. Also by performing small scale tests, we did not waste much material in the early testing phase. Although the small scale testing was more economical, the issue of changing scale of our fabrication became a challenge. All the factors previously mentioned such as - viscosity, nozzle diameter, temperature, air pressure, height above printing surface, are directly linked to the scale used. Consequently for each item that we had resolved, it had to be recalibrated as we moved up in scale. A lesson for future reference is to first decide what is the scale of the object to be 3D printed - food, art or construction, then tailor the fabrication method to suit.







TEMPERATURE This gelatin-based bioplastic becomes aqueous at temperatures above 30 degrees Celcius, and would develop surface tension below 25 degrees. Therefore in order to be able to 3D print this material, we needed to achieve this temperature range. Otherwise the material either become too liquid, or get blocked in the extruder. The aluminium extruder was heated by wrapping Nicrom wire at even intervals around it. This was insultated using Kapton tape, to prevent contact between the electrical current and the extruder, and also along the Nicrom wire itself. If any contact occurred along the circuit, it could create a short circuit. For the small extruder we used a 0.9mm wire at approximately 25mm spacing, this was connected to a 12V power supply with 1 Amp of current. The larger extruder required 4 x 12V power supplies, and 4 Amp, when using 10m of 0.6 Nicrom wire. We tested the temperature across the cylinder. Initially there would be a temperature variation in different locations. However over time the aluminium would redistribute the heat more evenly.


Primarily we would have a higher temperature of the extruder than the material itself, since there would be some heat loss during refilling, as well as from the exterior to the interior of the cylinder. This was tested manully each time. Also we found that it was best to begin with a heated bioplastic. Otherwise the material could set before we began 3D printing. To calculate the required resistance of the electrical circuit: R = Resistance I = Current V = Voltage P = Power R=V I R = V2 = P P I2




























05 In this chapter we will look at which computational tools we used to translate our ideas into a digitally designed and robotically fabricated fabricated prototype. At every stage of the design process we employed computational tools to further enhance our project; from the early concept and formfinding excercises, simulate material behaviour to structural analysis and the execution of the robotic toolpath.

evolve. It enables multiple options to be explored prior to materialization.

Rhinoceros was the main software used, as well as the algorithmic editor Grasshopper and its plug-ins such as; Kangaroo, Karamba and Kuka PRC, for various aspects of the project.

Another important aspect of the digital logics of this project since we were using robotic fabrication, was the use of Kuka|PRC. This plug-in to Grasshopper allowed us to program industrial robots directly from the parametric modelling environment. It also includes full kinematic simulation of the robot, which display any movement restrictions before the code is executed.

One of the advantages of using computational simulation, is to be able to generate and multiply design iterations to provide the designer and user with visual feedback on how the outcome might



In terms of structural analysis, parametric tools were employed to provide visual data about the structure. Based on this information of where stresses are present, the design could be adapted to suit. This is done preceding fabrication to minimise potential failure of the structure.



As a part of the digital logics for this project we developed a user interface, with the computational expertise of Rodrigo Aguirre. The idea behind the user interface was to create an easy, efficient way for the user to produce a global design of the structure and test different configurations of it. The logic of the user interface was based on the material system of the folding composite panels of the bioplastic. In our material investigation we learned that by increasing the percentage of food waste in the bioplastic, we could increase the degree of folding that was actuated. The Grasshopper plug in Kangaroo was used as an interactive physics solver, which would simulate and optimize the folding behaviour based on the input constraints. The user could input the available dimensions of their site, and how many panels they wanted to divide the structure into. One of the other parameters that could be input was the desired degrees of folding needed. From this input Kangaroo would find the optimum solution simulating physics spring systems. Part of the Grasshopper code generated would provide the resultant rest lengths of the geometry, to ease the fabrication process for the designer and user. This would provide the user with instantaneous graphical visual feedback to assist the in the decision making process.





Based on the global geometry generated in the previous formfinding excersice, this could then be analyzed using the parametric structural engineering tool Karamba. This tool is embedded in the design environment of Grasshopper. The inputs of Karamba include; global geometry of the shell/mesh structure, types of applied loads, support points, material, beams, cross sections thickness. From this input we can gain information about the force flows within the structure, displacement, utilization, peak stress points. Karamba can display this data visually and create an immidiate feedbacl loop to inform the designer where in the structure failure is likely to occur or where the structure is being under utilized, so potentially material could be removed from. In our particular design, the load case was a hanging structure where the maximum stress was concentrated around the top support points. Since we had this knowledge we could make an informed decision to have a higher density of food waste in those panels - given that we know that a higher content of cellulose would improve the structural performance of the bioplastic.


principle stress 1

It is interesting to note that the same region of the geometry that receives higher stress, is also where we have more food waste content to gain the folding and cellulose for strength. This demonstrates the direct link of form and function in biology, and natures way of finding minimal surface solutions.


principle stress 2



Initially we were generating the robotic toolpath based on 3D geometry modelled in Rhino. However we wanted to create a more robust and flexible system that could be reused several times with various input geometries. Our obiective was to vary the material deposition between the orange peel and coffee bioplastic in a gradient across the overall structure. Therefore each panel had a differential amount of the types of food waste embedded in them. From our prototype testing with robotic fabrication, we learned that with our particular material a continously 3D printed line worked better than discontinuity. In addition another aspect we learned during the additive manufacturing process was that intersecting lines in the same horizontal plane would often cause the material to disconnect. We discovered that with the bonding properties of the bioplastic material filleted corners created successful junctions. Another aspect of this Grasshopper definition was used to calculate how many grid divisions each neighbouring panel had, and create a transition between panels with a large spacing and smaller spacing.




Point of Flextion

ADDITIVE MANUFACTURING OVER TIME In addition to the system of composite panels we were creating, another challenge of our material investigation was to demonstrate that we could develop a material that was able to be 3D printed vertically. Due to the unique self-forming properties of the bioplastic, we wanted to work with this quality in mind when creating the vertical Vertical Stacking

toolpaths. Therefore we intentionally designed weaknesses into the structure, that would act as a point of flextion. The geometry of the 3D printed obejct would transform over time as a result of the curing process. We also learned by prototyping that the vertical step height in the z direction should be about 0.5 of the nozzle diameter. For example a nozzle diameter of 7mm, should have a step height of 3.5mm. This compression creates a bond with the previous layer, and therefore creates a more successful print.




Robotic Toolpath - One panel - All Panels

KUKA|PRC We used Kuka to simulate the range of robotic movement with the custom end effector for each 3D printed geometry, prior to executing the code. The Kuka-150 six-axis industrial robot was used for this project.


























The research performed to this point establishes a system to produce a material performance – the automatic formation of a rigid three-dimensional doubly curved bioplastic panel from a two-dimensional sheet cast – with a number of possibilities for further development. Having shown that a skeletonized panel exhibits the same contraction and automatic formation behavior as a solid panel, there exists significant potential for the employment of additive manufacturing techniques in the fabrication of lightweight, geometrically tuned panels. Printing the bioplastic rather than casting it would allow for precise control of the geometry, particularly if informed by a parametric model negotiating the transition from two dimensions to three. In particular, computer-controlled fabrication using a six-axis robot in concert with an air pressure extruder would potentially enable the fabrication of large-scale self-forming panels for architectural applications.

processes of metabolism and decay. In this sense, a construction material made from food waste has a role to play in urban strategy by diverting waste products away from landfills and into value-added goods. In digital materiality, there is an intrinsic link between technology, material science and organic form (Gramazio and Kohler 2008). Digitally fabricated self-forming panels bioplastic derived from food waste present the potential for design and industry alike to engineer functionally graded materials at low cost using abundant and largely untapped material resources. This research demonstrates the potential presented by food waste as a future sustainable building material.

However, the biological nature of bioplastic must not be forgotten. Organic matter decays as it biodegrades, presenting the challenge of designing materials that are strong and durable enough to be architecturally valuable, but not so robust so that they remain in the environment after they are no longer needed. Indeed, the ephemeral nature of biological materiality proposes an ephemeral architecture, subject to change over time under the






BAYER, Edward A., Henri Chanzy, Raphael Lamed, and Yuval Shoham. 1998. “Cellulose, cellulases and cellulosomes.” Current opinion in structural biology 8 (5): 548-557.

WOUDA, Hermanus A. J. 1981. Process for extracting ground roasted coffee. United States of America Patent 4,277,509. 7 July.

BROWNWELL, Blaine, 2012. Plastic: Material Strategies: Innovative Application in Architecture, New York: Princeton Architectural Press, p. 128-149 GENET, Marie, Alexia Stokes, Franck Salin, Slobodan B. Mickovski, Thierry Fourcaud, Jean-Francois Dumail, and Rens van Beek. 2005. “The influence of cellulose content of tensile strength in tree roots.” Plant and soil 278 (1-2): 1-9. GRAMAZIO, Fabio, and Matthias Kohler. 2008. Digital materiality in architecture. Baden: Lars Muller. HUERTA, Santiago. 2006. “Structural design in the work of Gaudi.” Architectural science review 49 (4): 324-339. KRETZER, Manuel, Jessica In, Joel Letkemann, and Tomasz Jaskiewicz. 2013. “Resinance: A (smart) material ecology.” ACADIA 13: Adaptive architecture. Cambridge: ACADIA. 137-146. RODRIGUEZ, Mauricio. 2012. “Bioplastics.” Materiabispent lity research network. 30 July.



WEBSITES reduce_drag_in_the/





Space Dynamics Workshop Manuel Kretzer

ACKNOWLEDGEMENTS This research was developed in the research studio Digital Matter|Intelligent Constructions. The group works with digital content, information and fabrication for the generation of new techniques generating the production of non-rigid, responsive and multi-functional material and construction systems. The authors would like to thank Areti Markopolou, Alexandre Dubor and Angelos Chronis, for their guidance and support during this project. In addtion, we thank our research colleagues and the IaaC faculty and staff.

Material Science Support Athanassia Athanasiou / Ilker Bayer IIT Istituto Italiano Di Technologia Robotic Fabrication Expert Djordje Stanojevic Additive Manufacturing Expert Sofoklis Giannakopoulous Computational Expert Rodrigo Aguirre Physical Computing Expert Angel MuĂąoz Coffee Supplier Orange Peel Supplier Cafe Cappucino Via Leitana 23, Barcelona




Profile for Lili Tayefi

Piel Vivo