Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978 Distributed Recycling of Waste Polymer into RepRap Feedstock Christian Baechler (1), Matthew DeVuono (1), and Joshua M. Pearce (2) 1. Department of Mechanical and Materials Engineering, Queen's University, Kingston, Ontario, Canada 2. Department of Materials Science & Engineering and Department of Electrical & Computer Engineering, Michigan Technological University, Houghton, MI 49931, USAAbstract:Purpose: A low-cost, open source, self-replicating rapid prototyper (RepRap) has been developed,which greatly expands the potential user base of rapid prototypers. The operating cost of the RepRapcan be further reduced using waste polymers as feedstock. Centralized recycling of polymers is oftenuneconomic and energy intensive due to transportation embodied energy. This paper provides a proofof concept for high-value recycling of waste polymers at distributed creation sites.Approach: Previous designs of waste plastic extruders (also known as RecycleBots) were evaluatedusing a weighted evaluation matrix. An updated design was completed and the description and analysisof the design is presented including component summary, testing procedures, a basic life cycle analysisand extrusion results. The filament was tested for consistency of density and diameter while quantifyingelectricity consumption.Findings: Filament was successfully extruded at an average rate of 90 mm/min and used to print parts.The filament averaged 2.805mm diameter with 87% of samples between 2.540mm and 3.081mm. Theaverage mass was 0.564 g/100mm length. Energy use was 0.06 kWh/m.Originality: This paper reports on the first technical evaluation of a feedstock filament for the RepRapfrom waste plastic material made in a distributed recycling device.Practical Implications: The success of the RecycleBot further reduces RepRap operating costs, whichenables distributed in-home, value added, plastic recycling. This has implications for municipal wastemanagement programs as in-home recycling could reduce cost and greenhouse gas emissionsassociated with waste collection and transportation as well as the environmental impact ofmanufacturing custom plastic parts.Keywords: Rapid prototyping; Open source; Polymer extrusion; Recycling; Waste plastic recycling;RepRapCategory: Technical1.0 IntroductionCommercial rapid prototypers allow for accurate fabrication of products or scale models, are a usefulproduction and design tool, and the development of additive manufacturing (AM) for rapid prototypingin a number of technologies has been substantial (Upcraft and Fletcher, 2003; Gibson, Rosen, and 1Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978Stucker, 2010; Petrovic, et al., 2010; Gebhardt, et al., 2010; Crane, et al., 2011). Recently an open source(OS) model, the RepRap, has been developed which can be built for under $1000, greatly expanding thepotential user base of rapid prototypers. The RepRap was created by Adrian Bowyer and influenced bymany contributors (Jones et al., 2011) including similar models such as the Fab@home (Malone andLipson, 2007). The RepRap is supported by an online wiki community with detailed assemblyinstructions for several variants of 3-D printers[1]. While OS models have limitations compared tocommercial processes, they are capable of creating highly accurate parts. The specifications of theRapMan, a commercially available version of the RepRap are shown in Table 1, below. These machinescould feasibly be used as a small-scale manufacturing or as an enabling tool for sustainabledevelopment (Pearce et al., 2010). Between 2008 and 2011, it is estimated that the number of RepRapsin use had increased from 4 to 4500 (Jones et al., 2011). Table 1: RapMan 3.1 Single 3D printing specifications[2] Parameter ToleranceBuild Size (x-axis / y-axis / z-axis) 270mm / 205mm / 210mmZ-axis resolution 0.125mmPrint Tolerance x axis and y axis: greater of ± 1& of object dimensions or ±0.2mm z axis: ± half the processed z resolutionPrint Speed Extruded Volume max 15mm3/s.Power Requirements 60W (5A at 12V)The operating cost of the RepRap can be further reduced by reducing the cost of the feedstock.Fabrication of feedstock from waste plastic would lower costs and reduce the environmental impact ofrapid prototyping (Pearce et al., 2010). Further, extrusion of RepRap filament may provide incentive fordistributed in-house recycling of plastic waste based on the preliminary economic analysis reportedhere. Currently, in-house recycling is marginal (e.g. using reclaimed objects in student art projects(TeacherVision, 2011)). The one growing exception is the household-scale recycling (composting) oforganic waste, which still tends to be centralized (Åberg, et al., 1996; Goldstein, 2003; EPA, 2011). Thecost and complexity of recycling most materials has historically favored centralized recycling withplastic recycling being a challenge for technical and social reasons (McDonald and Ball, 1998; Huang,et al. 2003; The Economist, 2007). Though favorable over alternatives, centralized recycling still hassome negative environmental impacts (Bjorklund and Finnveden, 2005; Ross and Evans, 2001). Inparticular, the collection, transport and transfer (CTT) of recyclable waste entails significantgreenhouse gas (GHG) emissions, contributing to the overall global warming potential of the recyclingprocess (Eisted et al., 2009; Salhoffer et al., 2006). In one specific case, transportation and collectionwas found to account for more than half of process emissions (Metcalfe, 2008).Waste plastic extrusion and OS 3D printing could also serve as a sustainable development tool (Pearceet al., 2010). Plastic waste is common in many developing communities (Al-Khatib et al., 2010;Muttamara et al., 1994) and informal waste recycling is sometimes conducted as an economic activity(Zia et al., 2008). A low-cost RecycleBot could serve to increase the product value of recycled waste inthese settings. A similar project is currently under way in Argentina known as Waste for Life whichprovides technology for value added recycling of waste polymers into plastic-fibre composite sheets 2Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978(Baillie, 2008). It should be noted, however, that before any sustainable development project isundertaken, an assessment of the social impacts of this technology should be undertaken to avoidunintentional impacts to environmental and social dynamics and assess the benefits to localpopulations.Commercial extrusion of plastic utilizes a screw to move material through a heated barrel where it iscompressed, melted, mixed and forced through a die (Rosato, 1997). Some groups have attempted toadapt these principles and construct a small-scale plastic extruder with varying success (RecycleBot,2010; Braanker et al., 2010). One such device, which turns waste plastic into a growth medium forplants (Torcellini, 2010), has been modified here to create a new, semi-automated RecycleBot to prepareRepRap feedstock from household plastic waste such as bottles and laundry detergent containers.A description and analysis of the semi-automated RecycleBot is presented here including componentsummary, testing procedures, a basic life cycle analysis and extrusion results. The filament was testedfor consistency of density and diameter while measuring electricity consumption. Results of thispreliminary proof of concept and technical evaluation of feedstock filament for the RepRap from wasteplastic material using a distributed recycling device are evaluated and conclusions are drawn.2.0 Methodology2.1 DesignThe design methodology incorporated examination of: i) commercial devices; ii) a literature reviewand; iii) similar open source projects. The following general desirable characteristics were identified forthe extruder: i) quality and reliability, ii) simple design minimizing number and complexity ofcomponents; iii) economical; iv) low power requirements; v) automated; and vi) the ability to acceptwide range of thermoplastics. Diametrical consistency (2.8±0.2mm) was needed to ensure the filamentcould feed consistently into the 3D printer, while constant density was necessary to support a steadyextrusion rate and high quality prints. A simple, low cost device will be more readily adopted by users.Maintaining low power requirements further increases the viability of the extruder in the Global Southand is environmentally beneficial. Automation will increase productivity while acceptance of a widerange of thermoplastics is required to meet all prototyping needs. Finally, the extruder must becompatible with the 3D printer, producing the plastic filament used as feedstock for the RepRap.Producing granules, or other feedstock would require alterations to the RepRap extruding mechanism,which is beyond the scope of this work.Three previous concepts were identified as a potential basis for the current design: 1) the RecycleBothand-crank extruder (RecycleBot, 2010), 2) Tan/Nixon granular extruder (Tan and Nixon, 2007) and 3)Torcellini's growth medium extruder (Torcellini, 2010). These design concepts were evaluated with aweighted evaluation matrix based on the following criteria: i) cost; ii) ease of use; iii) automation; iv)adaptability to RepRap; v) quality and; vi) energy use. An optimal concept was thus chosen, refinedand prototyped.Testing was conducted with high density polyethylene (HDPE). HDPE is characterized as relativelystiff with high tensile strength compared to other polyethylenes (American Chemical Council, 2007). 3Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978The engineering properties of HDPE are summarized in Table 2 (Khelif et al., 2008). Although otherpolymers exist within the household waste stream that are suitable for 3D printing, HDPE was chosento build on the work of similar open source projects, most of which used HDPE (Braanker et al. 2010,RecycleBot 2010, Torcellini, 2010). It was also found to be more prevalent in the local waste streamthan acrylonitrile butadiene styrene (ABS) (chemical formula: (C8H8)x· (C4H6)y·(C3H3N)z) andPoly(lactic acid) or polylactide (PLA) (chemical formula: C3H6O3), two polymers commonly used inOS 3D printing applications. Table 2: Common engineering properties of HDPE (Khelif et al., 2008). Property Value Density (kg/m ) 3 950 Thermal expansion factor (106/K) 100-200 Poisson’s Ratio 0.46 Friction factor 0.29 Rockwell Hardness (shore) D 60-73 Elastic modulus (GPA) 0.5-1.2 Impact Resistance, IZOD (J/m) 20-2102.2 TestingThe extruder prototype was tested with HDPE and the following metrics were quantified: i) resultantfilament consistency; ii) energy use per unit length of filament and iii) process time.Consistency was determined by measuring filament diameter and mass per unit length from a6.00±0.01m sample of extruded filament. The filament was divided into 60 sections of 0.100±0.003m.Diameter was measured at the section midpoint using a micrometer (±0.0001in/0.003mm). Sectionmass was determined using a digital scale (±0.0001g), and compared on the basis of mass per unitlength (±0.001g/100mm).The energy consumption was quantified using watt/electrical energy meters (±0.005 kWh). Data wasrecorded for each stage of filament production, including shredding (kWh/g), auger drive and heating.The energy use was measured for the extrusion of 11.41±0.05m of filament, and averaged to determinethe kWh/m of filament produced. Shredding energy was incorporated by converting the measure ofkWh/g to kWh/m as shown in equation 1 and 2 below. The energy density (energy per unit mass) ofshredded material, Q, is: Q = Es /mwp [kWh/g] Equation 1where Es is the energy consumed during shredding and mwp is the mass of the waste plastic shredded.The energy consumed during shredding per meter of filament produced Es/m is: Es/m = Q x ml [kWh/m] Equation 2where ml is the average mass per meter of filament.Process time includes the time required to shred, heat and extrude filament. An average time was 4Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978empirically determined for the shredding of 100g of polymer (min). This included removing labels,cutting bottles into smaller pieces and grinding using a commercial office shredder. It did not includethe time required for washing used bottles because of large variance in washing time based oncleanliness. Heating and extrusion time is divided into two components. The start-up time is defined asthe time passed between start-up and initial extrusion. During this time, the barrel must heat up andplastic remaining from previous extrusions must re-melt. The extrusion time is defined as the averagetime required to produce a meter of filament as timed with a digital watch (±0.01s).3.0 Results3.1 DesignThe Torcellini “aquaponics extruder” was modified and fabricated because of design strengthsincluding: automation, RepRap compatibility and product quality. The prototype waste plastic extruderconsists of 3/4" (inner diameter) piping divided into a gearing, hopper and heating section. Shreddedplastic is fed into the hopper and transported to the heating section by a 3/4” - 17" ship bore augerdriven by a windshield wiper motor. Power is transferred from motor to auger using a sprocket andchain drive system with a 2:1 gear ratio to decrease speed and increase torque. The motor has a stalltorque of 30Nm and runs at 35rpm on the low speed setting. The gearing section of the extruder bodyhouses the auger and also provides support for a thrust bearing and collar used to counter axial forceencountered during transportation of the ground plastic. An exploded assembly diagram of the mainextruder body is shown in Figure 1.Figure 1: Exploded assembly diagram of the extruder. Visible are the gearing, hopper and heatingsections, the auger, thrust bearing, collar and large sprocket.Plastic pushed into the heating section is melted and forced through a die producing plastic filament.The heating zone consists of 14 gauge nichrome 80 wire wrapped around the barrel and secured withheat resistant Kapton tape. Furnace cement was applied between the piping and the uninsulatednichrome to prevent short circuiting through the pipe at high temperatures. This step can be avoided byusing insulated nichrome wire. Fourteen gauge wire was chosen in an effort to create the heating zone 5Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978with minimal power requirements. Wire temperatures of 225oC were achieved with 75W of power (15V,5A). Temperature can be adjusted by altering the current passing through the nichrome wire. Thesection has one uniform heating zone. Presently, the heating section is uninsulated however addinginsulation would likely improve performance and reduce start-up time and energy use, while making itmore difficult to troubleshoot. Either a wood or metal base can be used to hold the extruder.3.2 Testing3.2.1 QualityDuring testing, 11.41±0.05m of filament was extruded. The first meter was removed because ofinconsistent extrusion on start-up. The remaining filament was divided into two sections. Qualitytesting was conducted on a 6.00m section of filament. The remaining 4.41m was used for printing.Measurements for mass and diameter were conducted on 60 samples of 100±3mm. A histogram ofdiameter measurements is shown in Figure 2, below. 16 12 Frequency 8 4 0 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3 Diameter [mm] Figure 2: Diameter measurements taken at 100mm intervals over 6m of filamentFilament diameter ranged from 2.225mm to 3.294mm with an average measurement of2.805±0.003mm.The mass of each section, displayed in Figure 3, ranged from 0.437g/100mm to 0.694g/100mm offilament. Measurements were distributed evenly across this spectrum. The average mass was0.564g/100mm. 6Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978 10 8 6 Frequency 4 2 0 0.45 0.49 0.53 0.57 0.61 0.65 0.69 0.43 0.47 0.51 0.55 0.59 0.63 0.67 0.71 Mass per Unit Length (g/mm) Figure 3: Mass of 100mm sections of filament from a 6m long sample.3.2.2 Energy ConsumptionThe peak power and measured energy consumption during each production stage is shown in Table 2.Shredding energy was calculated as an average of four trials producing 2060±10g of shredded plastic intotal and converted to energy per meter of filament using Equation 2. Heating accounted for two thirdsand the motor energy use about 1/3 of the total energy use, while shredding energy consumption wasnegligible. Table 3: Energy use at each stage of extruder process. Power Peak Power Energy per Percent of Draw [W] Length Energy per Filament Unit Length [kWh/m] [%] Shredding 128 0.0002 0.33 Motor 66 0.02 33.22 Heating 75 0.04 66.45 Total 269 0.06 100These values of energy consumption are necessary for a life cycle analysis (LCA), which is a means ofquantifying how much energy and raw material are used and how much (solid, liquid, and gaseous)waste is generated at each stage of a product’s life. Ideally an LCA would include quantification ofmaterial and energy needed for (Pearce and Lau, 2002): raw material extraction, manufacturing of all 7Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978components, use requirements, generation (if any -- e.g. photovoltaics), end of use (disposal orrecycling), and the distribution/transportation in between each stage. A full LCA is beyond the scope ofthis paper, however, a direct comparison between the recycled feedstock created by the processdescribed above and a virgin HDPE material can be made. As the mass per unit length of HDPE is5.64g/m a kg of HDPE is 177.3 m of filament. The total energy use for filament production (includingshredding, melting and extrusion) from Table 3 is 0.06 kWh/m, which results in an embodied energy of10.63 kWh/kg HDPE, which is 38.29 MJ/kg. In comparison, the average embodied energy of virginHDPE feedstock over 11 records is 79.67 MJ/kg (Hammond and Jones, 2008). It should be noted thatthis figure is HDPE material alone and there may be additional embodied energy for forming filament.It can thus be concluded that the recycled feedstock using the processes described here has a muchlower embodied energy than virgin material. A more detailed LCA that compares distributed recycledHDPE filament production to large-scale centralized recycling facility product over different populationdensities is left for future work. This more detailed study could look carefully at embodied energy andemissions, take into account transportation and disposal, which would be particularly relevant in thedeveloping world where recycling programs may be inefficient or nonexistent and virgin polymer maybe in short supply.3.2.3 TimeTime requirements were measured at each stage of the extrusion process. A total of 2060±10g of plasticwas shredded in 211min or 10.27min/100g. Given ml of extruded plastic, 100g of shredded plasticshould be sufficient to extrude ~17.7m of filament.For the extrusion process, the start-up time was found to be 45±2min when heating with 5A and 15V,but was reduced to 25min when supplying 10A, 15V. Filament extrusion was not constant; however, theaverage rate was 90mm/min.4.0 DiscussionSeveral parts were successfully printed on a 3D printer as proof of concept for the viability of the useof the recycled polymer filament.4.1 AnalysisFrom the initial samples, 65% was within the desired diametrical range of 2.28mm, while 87% fellbetween 2.540 mm and 3.081mm in diameter. This is less consistent than commercially extrudedfilaments such as ABS from Makerbot at 2.27±0.01mm. Inconsistencies measured here are a result of avariable extrusion rate and manual drawing of filament from the die.The mass per unit length of 60 samples was distributed between 0.437g/100mm and 0.694g/100mm, arange of 0.257g/100mm as can be seen in Figure 3. This range represents 46% of the average mass,found to be 0.564g/100mm. While the extruded filament worked when printing, this large variation cancause problems in print quality and further optimization of the system is required. The average ml is 8Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978close to expected values as HDPE of 950kg/m3 density should produce ml = 0.585g/100mm forfilament with a diameter of 2.8mm. The variation in mass can largely be explained by thecorresponding change in diameter between sections.The total electricity consumption required for filament construction was 0.060±0.05kWh/m of filament.Excluding labor costs and using current residential electricity prices in the U.S. at 11.2 cents/kW-hr(EIA, 2011) a kg of HDPE feedstock would cost less than a dollar to fabricate. For comparison, thecurrent cost of a kilogram of HDPE or ABS filament is between $37-$43.[3] These values are the retailrate of the filament, not the cost of production, which could be significantly lower as the 2011 U.S.prices for bulk purchase of polymer pellets ranged from $0.69/kg to $1.76/kg and $1.91/kg to $2.00/kgfor HDPE and ABS respectively.[4] It would be expected that the large-scale production of filamentcould have some energy efficiency benefits compared to the small systems described here, simply fromsurface to volume ratio heat loss reductions. Based on the raw material costs and the energy needs ofproduction, it appears clear that the cost of production are currently very low compared to retail costs.Although values for the cost of production of the filament using modern large high-speed extruders isnot available in the open literature, it is clear that in the current AM industry, the price of the filamenteven for the low-cost 3-D printers is significantly inflated. The widespread adoption of the RecycleBotsdescribed in this paper, should foster competition and a potential oversupply of filament on the market(e.g. RecycleBot owners could offer filament for sale on the web and the same time decreasing demandby providing for their own needs). Thus, except for rare exceptions (Chen and Riordan, 2008), just as inmany cases in economic history as the competition and supply of a manufactured good increases, thecosts can be expected to decrease. The existence of RecycleBots, therefore, should provide a significantdownward price pressure for conventionally manufactured filament at the retail level.The design performed well under most of its performance criteria. The extruder is simple to use, andconstruction of the device can be accomplished with basic metal working skills. The extruder cansuccessfully extrude HDPE with 125 W of power, which is about the same power draw as the shredder.With proper temperature control, a wide range of thermoplastics can be extruded but further testing isneeded to determine optimal settings for each.Three deficiencies were observed in the current design. First, physical assistance was required to drawthe filament from the extruder. This resulted in a wide variance of extrusion diameter due toinconsistent rates of pulling. Manually pulling also required constant user attention, reducing thebenefits of automation. Two devices were created in an attempt to automatically draw the filament at aconstant rate. One device used a motorized spool to coil the filament while the second employed apinch wheel design to pull filament from the die.These devices were ultimately unsuccessful due to the second extruder limitation - an inconsistent rateof extrusion. It was found that the rate of extrusion varied greatly over even short extrusion distances.This meant that neither device could not be calibrated to draw filament at an optimal rate, as theylacked the proper feedback mechanisms to adjust to changing extrusion speeds. At higher extrusionrates, it was possible to create a reasonably consistent product by manually drawing the filament.This rate variance was most likely due to the third limitation of heterogeneous waste feedstock andmethods of overcoming it include large batch mixing (blending) after shredding or extruding only from 9Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978homogeneous sources. Extrusion was greatly affected by the size and type of shredded plastic fed in tothe machine. Thin, light pieces of HDPE, such as those sourced from milk jugs, did not extrude well asthey were not easily drawn into the heating section of the extruder. Heavier pieces, sourced fromlaundry detergent and shampoo containers performed much better and extruded at a faster, moreconstant rate. Extrusion rate was further increased by cutting the feedstock into small bits (≤ 5mm x5mm), which greatly increased the extrusion rate, but did not improve consistency. Further work isnecessary to make the extrusion rate more consistent and fully automate the extrusion process. Onesuggestion is to extend the auger further into the heating section to have control over the movement ofplastic further into the process.Recycled plastic was successfully used in an OS 3D printer to make increasingly successful parts.Successive part generations showed increased accuracy, higher density and reduced delamination asextrusion rate, layer thickness and fill percentage were optimized. Thermal warping was also reduced astemperature settings were improved and surface finish was enhanced by adjusting settings for shelllayers. Parts created with recycled HDPE have not matched the quality of parts printed fromcommercially-available virgin ABS on the same machine. A number of factors contribute to thisdiscrepancy: i) HDPE is more difficult to work with than ABS because of it is more vulnerable to theeffects of thermal warping during printing[5], ii) ABS has been used on RepRaps and variants inthousands of locations for years, during which time the printing settings have been progressivelyoptimized, and iii) the variation in diameter and ml of filament make it difficult to print at a consistentrate. Initial experimentation with print settings has greatly contributed to an improved product and it isbelieved that further experimentation with these settings, as well as increased automation of futureRecycleBot variants, the quality of prints will continue to improve. The addition of a heated print bed isalso recommended to limit the effects of thermal warping on printed parts.4.2 ImplicationsThe success of this device further enhances RepRap affordability by reducing operating costssubstantially. This work therefore supports acceleration of the already rapid diffusion of RepRap, homemanufacturing, distributed manufacturing, “AM at home” or OS 3D printing concepts. It also reduceswaste associated with OS 3D printing as failed prints and excess filament can be recycled and reused inRepRaps. Research has been conducted into the degradation of crystalline additive manufacturingmaterials, but until now similar studies pertaining to HDPE and other amorphous polymers have beendeemed unnecessary, however with the growing prevalence of OS 3D printers, these materials must beconsidered.With a growing user base, shrinking cost and an increasingly user friendly interface, OS 3D printinghas major implications in small scale manufacturing and design. By further reducing operating costsand negating the need for shipping from a third party, recycled filament can further promote adoptionof OS 3D printers. Low price machines could make rapid prototyping more accessible for educationalpurposes, small businesses and domestic use.In addition, filament extrusion enables in-home, value-added plastic recycling thereby avoiding theGHGs and economic costs associated with municipal recycling programs by bypassing CTT. In theiranalysis of various recycling management systems, Salhofer et al. concluded that transport accountedfor up to 10% of process GHG emissions in HDPE film recycling (2006). Eisted et al. quantified CTT 10Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978as resulting in 9.4 to 368 kgCO2equiv./tonne of waste depending on various factors (2009). It should benoted that these studies are not identical materials to our product and that a more detailed LCA isneeded in the future. However, eliminating the need for collection and transportation of polymer wastewould clearly reduce environmental impact.5.0 Future WorkIn addition to the improvements needed to automate the RecycleBot with a constant rate of extrusiondiscussed above, research is needed to develop an inexpensive, open source household-scale plasticshredder. Plastic used in testing was shredded with a commercially available office shredder, which wastime consuming and created an inconsistent product.To show the viability of recycled feedstock for use in products which require high-quality materials, theengineering properties of recycled polymers should be evaluated and compared with those of virginpolymers and the impact of multiple extrusion cycles should also be examined as plastic is viable foronly a finite number of cycles (Karahaliou and Tarantili, 2009). These tests can also include the qualityof the recycled HDPE filament in terms of circularity, which is highly relevant for the RepRapdeposition process. No known testing has been conducted in the context of 3D printing.The potential of lower environmental impact of a distributed network of RecycleBots is based on theidea that this approach is more efficient than the use of large centralized (commercial) polymerrecycling facilities because of the substantial CCT demanded by centralized recycling. The necessaryembodied energy of transportation demanded by the centralized approach make it inherently inefficient.This CCT inefficiency is enhanced in the case of polymers because of the relatively low density whencompared to other recyclable materials. However, normally large-scale centralized processes tend to bemore efficient (both financially and in terms of energy consumption as mentioned above) due toeconomies of scale, whether in recycling or in production of even complex modern products such assolar photovoltaics (Prattern, 1971; O’Sullivan and Sheffrin, 2003; The Economist, 2007; Pearce,2008). Thus it is clear that a more thorough investigation is needed to determine if these economies ofscale exist in polymer recycling and whether they are outweighed by the advantages of localizedrecycling. As such, a full life cycle analysis of waste plastic feedstock as discussed above should beconducted and compared to centralized recycling, which accounts for multiple extrusion cycles andultimately, feedstock disposal. The results of this study will be geographically dependent as traveldistance, type of vehicle and fuel, and emissions intensity of the grid depend on the region investigated.This will give a better indication of the true environmental impacts of this technology. Finally, testsshould be conducted on the device to quantify the off-gassing of volatile organic compounds and otherharmful substances during plastic melting. This has been done for various commercial processes (Patelet al., 1995) and has found that worker exposure to harmful substances is below health and safetystandards (Forrest et al., 1995). However, the in-house process is much less controlled and should beindependently tested. 11Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/135525413113029786.0 ConclusionsA household-scale semi-automated waste plastic extruder was created by modifying existing designsand successfully extruded filament for use in an open source 3D printer. Recycled plastic from theextruder was successfully used in an OS 3D printer to make increasingly successful parts. Filament wasextruded at an average rate of 90 mm/min with average diameter of 2.8mm. While some productvariance occurred, 87% of samples had an acceptable diameter and the average mass of extrudedfilament was 0.564 g/100mm length. Energy use was 0.06 kWh/m, which is significantly lessembodied energy than virgin HDPE and makes filament extrusion at home roughly 40 times moreeconomical in terms of energy than purchasing filament commercially. By further reducing operatingcost, and negating the need for shipping from a third party, recycled filament can help grow the 3Dprinting community. It may also serve as an alternative source of income through value added recyclingin the third world and as motivation for in-house recycling of plastic waste.ReferencesÅberg, H., Dahlman, Sven, Shanahan, H., Säljö, R. (1996). Towards sound environmental behaviour: exploring household participation in waste management. Journal of Consumer Policy, vol. 19, pp. 45–67.Al-Khatib, I.A., Monou, M., Abu Zahra, A.S.F., Shaheen, H.Q. & Kassinos, D. (2010), "Solid waste characterization, quantification and management practices in developing countries. A case study:Nablus district – Palestine", Journal of environmental management, vol. 91 no. 5, pp. 1131-1138.American Chemistry Council, (2007) “Plastic Packaging Resins.” American Chemistry Council. Viewed: April 12, 2011. available at: http://www.americanchemistry.com/s_plastics/ bin.asp?CID=1102&DID=4645&DOC=FILE.PDF|available. (accessed 22 June 2011).Baillie, C. (2008), "Waste for Life: Student learning through international development projects. Who pays and who benefits?", Materials Today, vol. 11 no. 10, pp. 6.Bjorklund, A., and Finnveden, G., (2005), "Recycling Revisited-Life Cycle Comparisons of Global Warming Impact and Total Energy use of Waste Management Strategies," Resources, Conservation & Recycling, vol. 44 no. 4, pp. 309-317.Braanker, G. B., Duwel, J. E. P., Flohil, J. J., (2010), "Developing a Plastics Recycling Add-on for the RepRap 3D-Printer," available at: http://reprapdelft.files.wordpress.com/2010/04/reprap- granule-extruder-tudelft1.pdf, (accessed 22 June 2011).Chen, Y., Riordan, M.H. (2008). “Price-Increasing Competition”, The RAND Journal of Economics , vol. 39, no. 4, pp. 1042-1058. 12Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978Crane, N.B., Tuckerman, J., Nielson, G.N., (2011) "Self-assembly in additive manufacturing: opportunities and obstacles", Rapid Prototyping Journal, Vol. 17, Iss. 3, pp. 211-217.The Economist (2007). “Case history: The truth about recycling”, The Economist, Technology Quarterly: Q2 2007. Available: http://www.economist.com/node/9249262 (accessed Dec. 8, 2011).Eisted, R., Larsen, A. & Christensen, T. (2009), "Collection, transfer and transport of waste: accounting of greenhouse gases and global warming contribution", Waste Management & Research, vol. 27 no. 8, pp. 738-745.Energy Information Adminstration (EIA), (2011). “Average Retail Price of Electricity to Ultimate Customers by End-Use Sector, by State” Table 5.6. available at: http://www.eia.gov/cneaf/electricity/epm/table5_6_a.html, (accessed 23 June 2011).Environmental Protection Agency (EPA), (2011). “Wastes - Resource Conservation - Reduce, Reuse, Recycle – Composting” available at: http://www.epa.gov/osw/conserve/rrr/composting/index.htm (accessed 9 Dec 2011).Forrest, M. J., Jolly, A. M., Holding, S. R., (1995),"Emissions from Processing Thermoplastics", The Annals of Occupational Hygiene, vol. 39 no. 1, pp. 35-53.Gebhardt, A., Schmidt, F.M., Hötter, J.S., Sokalla,W., Sokalla, P. (2010) “Additive Manufacturing by selective laser melting the realizer desktop machine and its application for the dental industry”, Physics Procedia, vol. 5, Part B, pp. 543-549.Gibson, I., Rosen, D.W., Stucker, B. (2010) Additive Manufacturing Technologies: Rapid Prototyping to Direct Digital Manufacturing, Springer: New York.Goldstein, N. (2003) “Solid waste composting trends in the United States”, Biocycle,vol. 44, pp. 38-44.Hammond, G.P., Jones, C.I., 2008, “Embodied energy and carbon in construction materials,” Proceedings of the Institution of Civil Engineers - Energy, 161 (2), pp. 87-98.Huang, C.H.,Ting, C.Y., Chang, S.M., Reu, R. (2003). “Institutional innovation and pricing scheme for sustainable waste recycling in Taiwan,” In C.H. Liu, J.R. Yeh, and C.H. Huang eds., New Challenges for Sustainable Development in Millennia (Chung-Hua Institution for Economic Research, Taipei), pp. 279-318.Karahaliou, E. -., and Tarantili, P. A., (2009), "Stability of ABS Compounds Subjected to Repeated Cycles of Extrusion Processing," Polymer Engineering & Science, vol. 49 no. 11, pp. 2269- 2275. 13Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978Khelif, R., Chateauneuf, A. & Chaoui, K. (2008), "Statistical analysis of HDPE fatigue lifetime", Meccanica, vol. 43, no. 6, pp. 567-576.Jones, R., Haufe, P., Sells, E., (2011), "RepRap - the Replicating Rapid Prototyper," Robotica, vol. 29 no. 1, pp. 177-191.Malone, E., and Lipson, H., (2007), "Fab@Home: The Personal Desktop Fabricator Kit," Rapid Prototyping Journal, vol. 13 no. 4, pp. 245-255.McDonald, S., Ball, R. (1998). “Public participation in plastics recycling schemes”, Resources, Conservation and Recycling, vol. 22, no. 3-4, pp, 123-141.Metcalfe, P. (2008). "Audit of the Kerbside Recycling Services." The London Borough of Camden. Energy Audit Camden Report 3. ADAS, Wolverhampton, UK.Muttamara, S., Visvanathan, C. & Alwis, K.U. (1994), "Solid Waste Recycling And Reuse In Bangkok", Waste Management & Research, vol. 12, no. 2, pp. 151-163.Patel, S. H., and Xanthos, M., (1995),"Volatile Emissions during Thermoplastics processing—a Review", Advances in Polymer Technology, vol. 14 no. 1, pp. 67-77.Pearce, J. M. (2008). “Industrial Symbiosis for Very Large Scale Photovoltaic Manufacturing”, Renewable Energy, vol. 33, pp. 1101-1108.Pearce, J. and Lau, A. (2002). "Net Energy Analysis For Sustainable Energy Production From Silicon Based Solar Cells", Proceedings of American Society of Mechanical Engineers Solar 2002: Sunrise on the Reliable Energy Economy, editor R. Cambell-Howe, 2002.Pearce, J. M., Morris Blair, C., Laciak, K. J., Andrews, Nosrat, A., Zelenika-Zovko, I., (2010). “3-D Printing of Open Source Appropriate Technologies for Self-Directed Sustainable Development”, Journal of Sustainable Development, vol. 3 no. 4, pp. 17-29.Petrovic, V., Gonzalez, J.V.H., Ferrando, O.J., Gordillo, J.D., Jose Ramón Blasco Puchades, J.R.B., Griñan, L.P., (2010 ). “Additive layered manufacturing: sectors of industrial application shown through case studies”, International Journal of Production Research, vol. 49, Iss. 4, pp. 1061- 1079.Prattern, C.F. (1971). Economies of scale in manufacturing industry. Cambridge, England: University of Cambridge Press.RecycleBot Team. "RecycleBot," May-June 2010, RecycleBot blog, available at: http://recyclebot.tumblr.com/archive. (accessed 22 June 2011).Rosato, Dominick (1997). Plastics Processing Data Handbook (2nd Edition). Springer - Verlag. 14Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978Ross, S. & Evans, D. (2003), "The environmental effect of reusing and recycling a plastic-based packaging system", Journal of Cleaner Production, vol. 11 no. 5, pp. 561-571.Salhofer, S., Schneider, F., and Obersteiner, G., (2007), "The Ecological Relevance of Transport in Waste Disposal Systems in Western Europe," Waste Management, vol. 27 no. 8, pp. S47-S57.O’Sullivan, A., Sheffrin, S. M. (2003). Economics: Principles in action. Pearson:Prentice Hall.Tan, A., Nixon, T. (2007), "Rapid Prototype Manufacturing System", The University of Adelaide, Adelaide, Australia, available at: http://www.fabathome.org/wiki/uploads/c/cd/Rapid_Prototype_Manufacturing_System_- _Timothy_Nixon_&_Adrian_Tan_2007.pdf (accessed 23 June 2011).TeacherVision (2011), “Recycled Art” Available at: http://www.teachervision.fen.com/recycling/lesson- plan/3276.html (accessed Dec. 9, 2011).Torcellini, R., (2010), "Plastic extruder for growing media," . The Adventures of Bigelow Brook Farm blog. Available at:. http://web4deb.blogspot.com/2010/12/plastic-extruder-for-growing- media.html. (accessed 12 April 2011).Upcraft, S., Fletcher, R., (2003), "The Rapid Prototyping Technologies," Assembly Automation, vol. 23 no. 4, pp. 318-330.Zia, H., Devadas, V. & Shukla, S. (2008), "Assessing informal waste recycling in Kanpur City, India", Management of Environmental Quality, vol. 19, no. 5, pp. 597-612.End notes[1] http://reprap.org/wiki/Main_Page., Multiple contributors, RepRap wiki, Accessed: April 24, 2011.[2]http://www.bitsfrombytes.com/sites/www.bitsfrombytes.com/files/downloads/RapMan31_Specs_A4.pdf., RapMan 3.1 Technical Specifications, Bits from Bytes, Accessed: September 19, 2011[3] http://store.makerbot.com/plastic.html Plastic for Printing. Accessed June 22, 2011. http://www.bitsfrombytes.com/catalog/materials Materials. Accessed July 13, 2011.[4] http://www.worldscrap.com/modules/price/index.php World Scrap. 2011. “North AmericanRecycled Plastics Prices (US cents per pound)”. Accessed Dec. 9, 2011.[5] http://wiki.bitsfrombytes.com/index.php/HDPE Bits from Bytes wiki. Accessed: April 24, 2011. 15Christian Baechler, Matthew DeVuono, and Joshua M. Pearce, “Distributed Recycling of Waste Polymer into RepRap Feedstock” Rapid Prototyping Journal, 19(2), pp. 118-125 (2013). DOI: http://dx.doi.org/10.1108/13552541311302978AcknowledgementsThe authors would like to acknowledge A. Bryson and N. Alazzam for technical assistance.Christian BaechlerDepartment of Mechanical and Materials Engineering, Queen's University, Kingston, Ontario, Canadabaechler.christian@gmail.comChristian Baechler is a research associate at Queen's Applied Sustainability Research group, where his academic interestsinclude sustainable development and environmental applications of engineering.Matthew DeVuonomdevuono@gmail.comDepartment of Mechanical and Materials Engineering, Queen's University, Kingston, Ontario, CanadaMatthew DeVuono is a research associate at Queen's Applied Sustainability Research group, where his research focuses onopen source hardware design and sustainable development.Corresponding author:Joshua M. PearceDepartment of Materials Science & Engineering and Department of Electrical & Computer Engineering, MichiganTechnological University, Houghton, MI 49931, USApearce@mtu.eduMailing address: 601 M&M Building, 1400 Townsend Drive, Houghton, MI 49931-1295ph:906-487-1466Dr. Joshua M. Pearce received his Ph.D. in Materials Engineering from the Pennsylvania State University. He thendeveloped the first Sustainability program in the Pennsylvania State System of Higher Education as an assistant professor ofPhysics at Clarion University of Pennsylvania and helped develop the Applied Sustainability graduate program while atQueen's University, Canada where he headed the Queen’s Applied Sustainability Research Group. He currently is anAssociate Professor cross-appointed in the Department of Materials Science & Engineering and in the Department ofElectrical & Computer Engineering at the Michigan Technological University. His research concentrates on the use of opensource appropriate technology to find collaborative solutions to problems in sustainability and poverty reduction. 16