To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009 Life Cycle Analysis of Distributed Recycling of Post-consumer High Density Polyethylene for 3-D Printing Filament M. A. Kreigera, M. L. Muldera, A.G. Glovera and J. M. Pearcea,b,* a Department of Materials Science & Engineering b Department of Electrical & Computer Engineering Michigan Technological University 601 M&M Building 1400 Townsend Drive Houghton, Michigan 49931-1295* Corresponding authorph:1-906-487-1466 fax: 1-906-487-2934 email: pearce@mtu.eduAbstractThe growth of desktop 3-D printers is driving an interest in recycled 3-D printer filament to reducecosts of distributed production. Life cycle analysis studies were performed on the recycling of highdensity polyethylene into filament suitable for additive layer manufacturing with 3-D printers. Theconventional centralized recycling system for high population density and low population density rurallocations was compared to the proposed in home, distributed recycling system. This system wouldinvolve shredding and then producing filament with an open-source plastic extruder from post-consumer plastics and then printing the extruded filament into usable, value-added parts and productswith 3-D printers such as the open-source self replicating rapid prototyper, or RepRap. The embodiedenergy and carbon dioxide emissions were calculated for high density polyethylene recycling usingSimaPro 7.2 and the database EcoInvent v2.0. The results showed that distributed recycling uses lessembodied energy than the best-case scenario used for centralized recycling. For centralized recyclingin a low-density population case study involving substantial embodied energy use for transportationand collection these savings for distributed recycling were found to extend to over 80%. If thedistributed process is applied to the U.S. high density polyethylene currently recycled, more than 100million MJ of energy could be conserved per annum along with the concomitant significant reductionsin greenhouse gas emissions. It is concluded that with the open-source 3-D printing network expandingrapidly the potential for widespread adoption of in-home recycling of post-consumer plastic representsa novel path to a future of distributed manufacturing appropriate for both the developed and developingworld with lower environmental impacts than the current system.Keywords: distributed recycling; life cycle analysis; plastic; polymer; recycling; transportation energyList of AcronymsABS: acrylonitrile butadiene styreneCO2: carbon dioxideg: gramHDPE: high-density polyethylene1To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009kg: kilogramkWh: kilowatt-hourlbs: poundsLCA: life cycle analysisLDPE: low-density polyethyleneMJ: megajouleOS: open sourcePC: polycarbonatePET: polyethylene terephthalatePP: polypropylenePS: polystyrenePVC: polyvinyl chlorideRepRap : self replicating rapid prototyperRecycleBot : post-consumer plastic extrudert: ton1. Introduction Plastic has become an integral part of society as population growth and technologicaldevelopment have resulted in the global production of plastic increasing by 500% over the last 30 yearsand it is expected to continue to grow to 850 million tons per year by 2050 (Lebreton, et al., 2012;Lotfi, 2009; Shen et al., 2009). In addition to food packaging and cheap parts, plastics are now beingused to replace metal, wood, paper, and glass in a variety of engineering applications (Arena et al.,2003), mulch (McCraw and Motes, 2012), sports fields, and even human body parts (Bow and Parvizi,2011; Xue, 2011). This increase in plastic usage results in a substantial environmental burden on bothland (Rees, 1980) and water pollution (Derraik, 2002) as plastics are slow to decompose naturally –taking from 10 to 450 years in a landfill (U.S. National Park Service, 2012) and toxic to burn (Lewisand Sullivan, 1992). Plastic processing, use, and disposal also comprise a significant source of energyconsumption (Björklund, 2005; Rydberg, 1995; Song and Hyun, 1999). This has been largelydetermined from life cycle analysis studies of plastic (Arena et al., 2003; Reich, 2005) and recycling(Craighill, 1995; Perugini et. al., 2005; Powell, 2010; Ross, 2003; Subramanian, 2000). First, plasticscan be regarded as a form of stored potential energy as each year producing virgin plastics requires 4%of the world’s oil production (Cambridge-MIT Institute, 2005) equivalent to 1.3 billion barrels a year(U.S. EIA, 2011) equivalent to the amount of oil Texas used in 2010 (U.S. EIA, 2010). As the cost ofoil is expected to escalate due to rising energy prices it is likely companies will look for alternativefeedstocks (Chemical Engineering Progress, 2008), thus there are both strong environmental as well aseconomic interests in large-scale recycling of plastics (Lotfi, 2009). Today seven types of plastics are commonly recycled including: 1) polyethylene terephthalate(PET), 2) high-density polyethylene (HDPE), 3) polyvinyl chloride (PVC), 4) low-density polyethylene(LDPE), 5) polypropylene (PP), 6) polystyrene (PS), and 7) “other”, which is primarily polycarbonate(PC) and acrylonitrile butadiene styrene (ABS). Both primary and secondary recycling schemes arewell established and widely applied (Al-Salem et al., 2009). Historical trends in polymer recyclinghave been towards large centralized facilities to take advantage of economies of scale in producing2To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009low-value commodities (Missouri DNR, 2012; Redd, 1993). One of the primary reasons that plastichas historically been recycled at very low rates (e.g. 6.5% in 2008) in conventional centralizedrecycling in the U.S., is the challenge of collection and transportation for high volume, low weightpolymers (Themelis et al., 2011) . Thus, plastic recycling is often not economical and when it isrecycled, the collection is ‘subsidized’ by higher value recycled content material such as aluminum(Hood, 1995). Two recent open-source hardware technological developments, 3-D printers and RecycleBots,offer a new approach to polymer recycling encompassing the potential for distributed processing tohigh-value added products, which reverses the historical trend towards centralized recycling facilities.Commercial 3-D printers, which allow for accurate fabrication of products or scale models, are a usefulproduction and design tool. 3-D printers allow for additive manufacturing, which is a process by whichdigital 3-D designs are used to build a component in layers by depositing material (Crane et al., 2011;Gebhardt et al., 2010; Gibson et. al., 2010). This is to differentiate it from subtractive (and thusnormally more wasteful methods) or conventional polymer manufacturing methods like plasticinjection molding (Peças et al., 2009). The development of additive manufacturing for rapidprototyping and 3-D printing in a number of technologies has been substantial (Petrovic et al., 2010;Upcraft and Fletcher, 2003). Recently, following the open source (OS) model, the RepRap has beendeveloped that can be built for under $500, greatly expanding the potential user base of 3-D printers.Between 2008 and 2011, it is estimated that the number of RepRaps in use had increased from 4 to4500 (Sells et al., 2011). These machines could feasibly be used for small-scale manufacturing or as anenabling tool for green manufacturing (Kreiger and Pearce, 2013a; 2013b; Pearce, et al., 2010). Theprimary expense of operating a 3-D printer is the filament or “3-D ink” and thus the operating costs ofthe RepRap can be further reduced using post-consumer plastics as feedstock. Commercial extrusion of plastic utilizes a screw to move material through a heated barrel whereit is compressed, melted, mixed and forced through a die (Rosato, 1997). One such device, which turnspost-consumer plastic into a growth medium for plants (Torcellini, 2010), has been modified here tocreate a new, semi-automated open source “Recyclebot” to prepare RepRap feedstock from post-consumer household plastic, such as, bottles and laundry detergent containers (Baechler et al., 2013).Researchers throughout the world have attempted to adapt these principles and construct small-scaleplastic extruders with varying degrees of success (Braanker et al., 2010; Kreiger et al., 2013;RecycleBot, 2010). As the RecycleBot is an open-source project there are several other variants underdevelopment: the Filabot, which includes an open-source shredder (McNaney, 2012), the LymanFilament Extruder (Lyman, 2012) and the MiniRecycleBot (MiniRecyclebot, 2012), which could beutilized as post-consumer plastic RecycleBots. Bad prints, broken or worn out parts can also berecycled by this method. There is also work currently being done in the open-source community on thecreation of a shredder designed to be used with the RecycleBot system. These open-source shreddersare capable of shredding entire milk jugs or other recyclables. Their use, in place of a commercial papershredder, would remove the need for cutting bottles by hand, thus reducing processing time (Thymark,2012). The use of an open-source shredder would also increase the usable mass of post-consumerplastic containers. Fabrication of feedstock with RecycleBots from post-consumer plastic also has the potential toreduce the environmental impact of 3-D printing, and may provide an incentive for distributed, in-3To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009house recycling of plastic (Pearce et al., 2010). Baechler et al. (2013) have demonstrated acceptable 3-D filament production from a RecycleBot using high density polyethylene (HDPE). HDPE, recycledplastic number “2”, is used primarily for non-food packaging, pipes, and plastic lumber. In 2010, 984million pounds of HDPE were recycled in the U.S. (American Chemistry Council and Association ofPostconsumer Plastic Recyclers, 2011), which is 27.8% of the HDPE produced (Sandhill Plastic, 2010).The total energy used annually to produce virgin HDPE is 124 billion MJ. Grant and James (2005)showed the embodied energy of recycled HDPE is 24% less than virgin plastic using their cut-offmethod of life cycle analysis (LCA), indicating that in there is an enormous waste of energy andmaterial resources from not recycling HDPE. This study explores this technical potential of using a distributed network of RecycleBots to processpost-consumer goods into 3-D printing feedstock. To demonstrate the feasibility of this approach,HDPE is used as a test material. The LCA of energy consumption and carbon dioxide (CO2) emissionsis determined for this distributed approach and are compared to the standard centralized model. Asensitivity analysis is performed on two case studies comparing both the best and worst case scenariosin Michigan of geographic distribution of post-consumer plastic from a centralized facility. Theseresults are discussed and a preliminary financial analysis is performed to draw conclusions about theviability of distributed recycling.2. Methods The goal of this study is to determine whether the use of a RecycleBot to create 3-D printingfilament using HDPE is an environmentally feasible alternative to conventionally recycling HDPE.The cases looked at were distributed recycling using the RecycleBot and three cases for conventionalrecycling: highly populated area, low population area with biweekly recycling trips, and low populationarea with monthly trips for recycling. The other cases were then compared to the literature value ofaverage embodied energy of virgin HDPE feedstock (Hammond and Jones, 2008). The scope of thislife cycle analysis will be limited to inputs for the processing due to recycling using the RecycleBotversus conventional recycling. The functional unit is 1 kg of recycled HDPE usable for either 3-Dprinting or conventional manufacturing processes. The inventory data associated with these inputs willembody a “gate-to-gate” system boundary, with the gate starting at the end of first useful life of theHDPE within the consumer’s home and ending immediately after production of a recycled filament orpellet. These results will be compared and used with previous LCA results on HDPE recycling fromliterature, to quantify the difference in energy demand and greenhouse gas emissions. SimaPro v 7.2 inconjunction with the EcoInvent v 2.2 database of materials, was used to complete an energy andemissions LCA of a distributed recycling (RecycleBot) versus conventional recycling. Cumulativeenergy demand (CED) was used to analyze the overall energy costs and the model developed byIntergovernmental Panel on Climate Change in 2007 for the global warming potential over a 100-yeartime period, IPCC 2007 GWP 100a, was used to calculate the CO2 equivalent emissions for therecycling comparison (Pachauri & Reisinger, 2007; Hischier, et al., 2010). The LCA began with collection and transportation of post-consumer plastic through therecycling process as indicated for the conventional recycling process shown in Figure 1. For thedistributed recycling process, the LCA was calculated for the plastics from transportation throughfilament drawing in the RecycleBot as seen in Figure 2. LCAs were completed for both processes usinga best and worst case scenario outlining the maximum and minimum collection distances in the U.S.4To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009 As shown in Figure 2, for the distributed recycling case post-consumer plastic is first collectedin the home and cut with scissors to be fed into a commercial paper shredder. The mass of the usableplastic was measured on a digital scale (±0.05 g). This method produces 49g of usable mass (87% oftotal mass) of a milk jug, thus 20 jugs/kg. These shreds are then fed into the post-consumer plasticextruder (RecycleBot)1 , which melts them and forms a 3mm filament that is then used in a RepRap orother 3-D printer, as detailed in Baechler et al. (2013) and Kreiger, et al. (2013). The RecycleBot usedin this study had improved insulation (1.5 inch high-temperature calcium silicate wrapped in kaptontape) compared to no insulation (Baechler et al., 2013) and modest insulation (Kreiger, et al, 2013). The energy consumption of the distributed recycling process was quantified experimentallyusing a multimeter (±0.005 kWh) that monitored an insulated RecycleBot during extrusion of 10m andaveraged. Data was recorded for each stage of filament production, including shredding (kWh/g), augerdrive, heating and a filament spooler, however previous work showed that the shredding even in anuninsulated RecycleBot was negligible and was ignored here (Baechler et al., 2013). The experimentalvalue was then put into SimaPro using the input (Electricity, Production Mix, US). In the conventional recycling process (Figure 1), after a plastic product is purchased and used ifit is recycled it is collected at curbside and transported to a collection center. At the collection center,HDPE is sorted to produce bales. After separation the HDPE bales are sent to a reclamation facility tobe purified and pelletized to be sold to manufacturers to create new goods. An example of a “best case scenario” for conventional recycling is a city like Detroit, Michigan,where there are four collection centers in metro Detroit, as well as curbside pickup, which funnel intoone processing center (Horton, 2009) for greater Detroit, as shown in Figure 3. After separation at theprocessing center, the plastic bales are sent to BATA plastics in Grand Rapids, Michigan, 157 milesaway, where the bales are made into quality pellets and sold to manufacturers for re-use. It is assumedthat the curbside recycling trucks are never at capacity and would make the curbside pickup without thecollection of post-consumer plastic and that the additional weight of the plastic would have a negligibleeffect on the fuel efficiency of the collection vehicles. Therefore only the embodied energy oftransportation and emissions due to the transport from the collection centers to the processing centerwere included and normalized per unit mass. A life-cycle inventory study of conventional recycling of HDPE was previously completedusing confidential information from recycling companies to quantify the impact and energy demand(Franklin Associates, 2011) and was used as an approximation here. This literature evaluated therecycling system in California using confidential information from material recovery facilities, plasticrecycling facilities, and HDPE reclaimers. Aspects not included in the literature consist of capitalequipment, space conditioning, support personnel requirements, and miscellaneous materials andadditives and thus will not be used in this study either. The values used for quantifying theconventional recycling were taken from the literature value for “cut-off weight based” result for therecycling of post-consumer HDPE, this method does not take the original burden of the first life of theproduct into account and does calculations by weight instead of volume. The value for the cumulativeenergy demand was 3.87 MMBtu per 1000 lbs. As this value was originally MMBtu for 1000 lbs, itwas converted to kilograms then divided to represent 1 kg, and then the result was converted fromMMBtu to MJ. A similar unit conversion was also done for the greenhouse gas emissions. Thesevalues are for biweekly recycling drop-offs.1http://www.thingiverse.com/thing:129485To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009 For the “worst case scenario” for centralized recycling, which represents a lowpopulation density, small, geographically isolated town of Copper Harbor, Michigan was used.Located at the top of the Keweenaw Peninsula, Copper Harbor is a 48 mile drive north of the nearestrecycling collection center in Houghton, Michigan and there is no curbside pickup. From Houghton,Michigan the plastics are then driven in a garbage truck to Green Bay, Wisconsin, 212 miles away, tothe processing center. The average household generates 16.9 pounds of recyclables per week (U.S. EPA, 2010). The average amount of HDPE post-consumer waste, out of the total amount of recyclables, isapproximately 5.2 pounds or 31 % (Franklin Associates, 2011). For this case, two options areconsidered, recycling biweekly or monthly. The inputs used in SimaPro for conventional recyclingwere (Operation, passenger car, petrol, fleet average, 2010, Switzerland) for the round trip drive fromCopper Harbor to Houghton, Michigan, (Operation, lorry 3.5-20t, full, fleet average, Switzerland) forthe drive from Houghton to Green Bay, Wisconsin, and a similar input for an empty truck for the returntrip to Houghton. Assuming the average load for the truck was 38,990 lbs and a contamination amountof 8% (Franklin Associates, 2011). The energy was divided out based on the mass (kg) hauled, underthe conservative assumption that hauling the recyclable materials does not change the fuel efficiency ofthe vehicle. The result from the person-km is multiplied by the fraction of HDPE over the total weightof recyclable materials in the trunk (16.9lbs/wk). This result is multiplied to provide 1kg of HDPE +8% waste so the end result is MJ/kg useable HDPE. To summarize: the "person" in the person-km is"one trunk load" and then the value adjusted to account for how much of that trunk is allocated toHDPE, and finally how many of those portions are necessary to obtain the functional unit of “1kg ofrecycled HDPE”. The life-cycle inventory study of conventional recycling of HDPE (FranklinAssociates, 2011) was used again as an approximation here for the processing burdens. Similarly to the“best-case” conventional recycling, the values used for quantifying the “worst-case” conventionalrecycling were taken from the literature values for the “cut-off weight based” result for the recycling ofpost-consumer HDPE. The value used for this scenario was for processing only, leaving 7.51 MJ perkg HDPE recycled. From here, the collection and transportation amount calculated in this study wasadded onto the process amount to achieve the total amount of cumulative energy demand. This wasthen done similarly for the greenhouse gas emissions.3. Results Table 1 summarizes the results of the embodied energy and greenhouse gas emissions forcentralized and distributed recycling of HDPE in the high population density and low populationdensity cases. The embodied energy values can be compared directly. The RecycleBot required an initial heating provided by 0.06 kWh before starting any extrusionand a running requirement of 0.0036 kWh per meter of filament produced. The initial heating amount isapplied to the production of 1 kg filament output and is assumed to overburden this process; thisamount would realistically be allocated over the entire amount produced. As the mass per unit length ofHDPE is 5.64g/m, a kg of HDPE is 177.3 m of filament. The total energy use for filament production(including shredding, melting and extrusion) is 0.694 kWh per kg HDPE filament, which is about 2.5MJ/kg. In comparison, the average embodied energy of virgin HDPE feedstock is 79.67 MJ/kg6To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009(Hammond and Jones, 2008). It should be noted that this figure for conventional processing is HDPEmaterial alone and there may be additional embodied energy for forming filament acceptable for 3-Dprinting. Despite this it is clear from the results that using distributed recycling reduces embodiedenergy of HDPE over virgin material by 89%.The RecycleBot's greenhouse gas emissions werecalculated using SimaPro IPCC 2007 GWP for 100 years, with the only input being the electricityproduction mix of the U.S. It should be pointed out here that extreme care is necessary in comparingthe greenhouse gas emissions shown in Table 1. The kg CO2 eq per kg HDPE is heavily dependent onthe emission intensity of the grid where the case is run. Previous work by Kreiger et al. showed thatusing low-emission intensity solar photovoltaic devices for distributed electrical generation fordistributed recycling significantly reduces overall emissions for HDPE filament fabrication (2013). There is also an apparent discrepancy between the energy demand and GHG emissions forrecycled and virgin resins; as the former differ by an order of magnitude and the latter by a few percent.The reason for the difference in magnitude between the energy demand and the emissions is that therecycled HDPE does not include the energy content of the materials that end up in the product, but theenergy content of any fuels or electricity inputs to the recycling process are included. Whereas, thevirgin HDPE counts the energy content of the resources from nature (e.g crude oil and natural gas) thatgo directly into the HDPE plus the energy content of any fuels or electricity inputs used for theconversion processes. Since the natural resources used to make virgin HDPE are not used as fuel, theemissions are not released, thus the difference in magnitude. Table 1: Energy Demand & Greenhouse Gas Emissions Percent Reduction Greenhouse Gas Emissions Energy Demand Case (%) for Distributed (kg CO eq per kg HDPE) (MJ/kg HDPE) 2 Recycling c Distributed Recycling: 8.74 -- 0.52 Insulated RecycleBot Virgin Resin 79.67a 89 1.82b Centralized Recycling – High Density Popula- 9b 3 0.63b tion: Detroit Centralized Recycling – Low Density Popula- 28.4 69 2.65 tion: Copper Harbor (monthly) Centralized Recycling – Low Density Popula- 48.9 82 4.04 tion: Copper Harbor (bi-weekly) Notes: a (Hammond and Jones, 2008) b Estimate based on (Franklin Associates, 2011) c. Percent reduction = (Central-Distributed)/Central*1007To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009 The amount of energy for the conventional cases attributes 7.51 MJ eq from the recyclingprocess and transport needed between each step (Franklin Associates, 2011), while the remainder isfrom transportation due to collection and from the collection center to the processing center. As can beseen in Table 1, when comparing the centralized recycling for low density population on either a bi-weekly or monthly recycling trip case, the distributed recycling can decrease the embodied energy by69%-82%. In addition, even with varying emission intensities it is clear the distributed recycling isbeneficial from an ecological standpoint. In these low population density cases the embodied energyand emissions for the personal transportation of the HDPE to a collection facility have a substantialimpact on the LCA values and is added to the values for the complete process in the high-populationdensity case. Every 2 weeks every kg of HDPE consumes 41 MJ from the round-trip drive from CopperHarbor to Houghton, Michigan. This amount of energy for simple collection dwarfs the entireembodied energy of a high population density centralized recycling or the distribute recycling cases.The use of conventional recycling in this rural case is worse for the environment in terms of energy useand emissions than creating all new products from virgin resin. For such locations, these values oftransport can only be reduced by transporting more recyclable materials per trip. Thus, if one month ofthe total amount of post-consumer recyclables could fit in the vehicle, this results in an energyconsumption of 20.5MJ/kg HDPE for collection and thus a total embodied energy of only 28.4 MJ/kg,which is again about one third of that of virgin material.4. Discussion The results clearly show that distributed recycling of HDPE uses less energy than conventionalrecycling. If the population density is spread out these reductions can be significant ~70% reduction,but even in the best case scenario for conventional recycling it still uses 3% more energy thandistributed recycling. A 3% reduction should be looked at with caution as there is a 0.5% error inexperimental measurement and a small improvement in conventional recycling can outweigh thisamount for the ideal case. It should be noted that a switch to renewable energy can reduce thesenumbers. If the 984 million pounds of HDPE that are currently recycled in the U.S. per year in the bestcase for conventional recycling are instead diverted to distributed recycling, would result in the savingof 116,000,000 MJ of energy, equivalent to the energy used by more than 2,800 American household'sper year (U.S. EIA, 2011). However, it can be presumed that as the economic value of producing 3-Dfilament from household recycling became widespread the recycling rate could also be increased fromthe current value of less than a third. If the total HDPE supply was recycled using the distributedprocess, offsetting all virgin HDPE, over 100 billion MJ could be conserved per year. This study is a conservative estimate of the benefits of distributed recycling, because thecommercial pellets used for the comparison of the centralized case would then have to be furtherformed into filament. However, the same raw feedstock could be used directly to make filament ratherthan pelletizing (e.g. the same polymer melter/auger machine could be used to make filament with aspooler). The energy associated with the spooler or the pellet cutter is negligible when compared tomelting the plastic (Baechler et al., 2013). It was therefore conservatively assumed the two processesare equivalent. The distributed recycling advantage comes from both the low-overhead equipment cost of the8To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009RecycleBot and the complete elimination of embodied energy for transportation, which, as can be seenin the Copper Harbor examples, can be substantial. It is clear that recycling of HDPE should occur asclose to the source of plastic waste as possible with the largest load and lowest fuel consumption toreduce environmental impact. The results for both embodied energy and emissions show that for caseswhere the user is far from a recycling center, it is significantly better for the environment to do on-siterecycling using a RecycleBot, than to use conventional recycling of any kind. If close to a recyclingcenter, on-site recycling reduces energy demand and will reduce emissions as well. Not only doesdistributed recycling reduce energy and emissions, it will have farther reaching implications. The amount of fossil fuels and concomitant GHG emissions required to transport billions ofpounds of plastics from residences to collection centers to processing centers is considerable. If allplastics, which make of 10.5% of all recycled goods (Ohio DNR, 2012) are largely eliminated from thewaste stream these trucks would not have to circulate as often. Eliminating even a fraction (e.g. 3% viaelimination of the HDPE) of that through distributed recycled would be environmentally beneficial.The decrease in materials would be substantial as well. With fewer waste management trucks travelingover public roads, less damage will be done to the roads, requiring less frequent road maintenance.Similarities, when fewer plastics go through the large scale recycling system, less material is used tobuild and maintain the facilities. Combining the open-source distributed recycling of the RecycleBot with the distributedproduction of the RepRap combination systems would be the most economically beneficial for thoseinterested in a complete distributed manufacturing process. This could even be accomplished on ahousehold level. The RecycleBot could be used for disposing of a single household’s recycling, savingtrips to return waste plastic and a stop for curbside collection. The RepRap could be used to print partsfor simple household repairs and solutions, such as bike parts, knobs and handles, cooking utensils,toys, eyeglass frames, and an enormous sum of individual parts. Although these parts or products canoften be purchased in most locations, they can be printed often for considerably lower costs and bemade more customized or appropriate for the consumer (Wittbrodt, et al., 2013). For example, at anaverage U.S. utility cost of $0.1153/kWh (Electric Choice, 2010), an orange juicer of volume 63.4 cm3of PLA uses 0.31 kWh to produce on a RepRap (Kreiger and Pearce, 2013a; 2013b). This would costabout 3.5 cents in electricity to print. If purchasing commercial filament at $36/kg of filament for a75.47g juicer of 3 mm PLA filament, this costs $2.72 in material, for a total cost of $2.76. Citrus juicersof lower quality and utility can be found on-line for $1 and of approximately the same quality at $7-$25, so distributed manufacturing is not necessarily economical for low-value plastic products.However, if a RecycleBot is used to make the juicer from HDPE the material costs drop to theelectricity needed (0.56 cents), bringing a total manufacturing cost to a remarkably low 4 cents. Thusthe RecycleBot/RepRap combination allow for two orders of magnitude price decreases even for low-value mass-produced plastic products if people are willing to invest their time to make them. As haspreviously been shown, such radical decreases in costs can be obtained for high-value items such asscientific equipment (Pearce, 2012;2014; Zhang, et al., 2013) using only the RepRap and would bereduced even further using recycled post-consumer plastic for filament as shown in this paper. Thepotential for high-value added recycling using this process is already being investigated in thedeveloping world (Feeley et al., 2014) with the creation of the Ethical Filament Foundation to helpenable waste pickers to make 3-D printer filament.2 Future work is necessary to investigate the2http://www.ethicalfilament.org/9To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009willingness of consumers to 1) accept 3-D printed plastic goods, 2) recycled filament, and 3) potentialof the majority of consumers to become producers. There is a financial incentive for RecycleBot operators to produce filament even withoutmaking a finished product. The cost of commercial 3-D filament (ABS or PLA) currently ranges from$36-50/kg, while the RecycleBot iproduces 1 kg of filament from about 20 milk jugs for under 10 UScents. The RecycleBot operator could also produce pellets trading in a spooler/winder for a cutter andsell recycled HDPE on the open market for centralized manufacturers to use to make products such aswood plastic composite decking (Bolin and Smith, 2011). Although technically possible it is noteconomically advantageous for home recyclers to invest their time in this way as pellets cost ~$1/kg,while 3-D printer filament is currently selling at ~$35/kg and with a RepRap the recyclers can printproducts using the filament worth $100s-$1000s/kg (Wittbrodt, et al., 2013). Those involved with the 'maker movement' or the more traditional informal economy wouldalso benefit financially from this distributed manufacturing/recycling process. This sector is made upof people who are self-employed, untaxed, and unmonitored by the government, but still work withinthe legal limit (Sparks and Barnett, 2010). These people often run small-scale service businesses thatwould be well suited for a rapid prototyper. For example, someone who does computer repairs out oftheir home could use a RepRap printer to print any of the following: computer mouse, computer case,I/O cover plates; adapter brackets for hard drives and SSD, laptop stand, laptop privacy shields,wireless chording keyboard, docking stations, and keyboard parts.3 Thingiverse.com, which is arepository of digital designs most of which can be printed on a RepRap, currently holds over 200,000open-source designs and is growing exponentially (Wittbrodt, et al., 2013). There is a network effect tosharing open-source hardware designs as each design added to the commons adds value to all existing3-D printers. The low cost and reproducibility of the RecycleBot and RepRap and their product designs makethem useful in a developing world (Pearce, et al., 2010). A self-replicating 3-D printer can be used tomake appropriate technologies for energy generation, water distribution, utensils, shoe insoles, parts ofmedical equipment, parts of water filters, etc., as well as spare parts and copies of itself (Pearce, et al.,2010). A major concern in the developing world is water distribution. One of the Peace Corpsobjectives is to install drip irrigation to places where water is limited, as drip irrigation efficiently useswater without much loss to evaporation (Peace Corps, 2011). 3-D printers can be used to make partsand fittings for drip irrigation, potentially changing the water shortage in developing countries andsolving the food crisis in many areas.4 The ability to make these useful parts or products from recycledwaste using the distributed recycling paradigm discussed here, not only has the potential to radicallyreduce the environmental impact of HDPE-containing products, but also to substantially reduce costsfor developing world communities. Lastly, it should be noted that similar potential exists for otherrecycled plastics such as ABS, PLA, and PET and further work is needed to perform both technicalanalysis and LCAs on these materials in a distributed recycling process as they provide more easilyprinted 3-D filament and may be more likely to be accepted by 3-D printer operators.3 Examples from http://www.thingiverse.com/thing:[#] computer mouse [1056] , computer case [3944], I/O cover plates[14377]; adapter brackets for hard drives and SSD [13472], laptop stand [7346], laptop privacy shields [8412], wirelesschording keyboard [6922], docking stations [16608], keyboard parts [13015]4 See examples http://www.thingiverse.com/jpearce/collections/open-source-appropriate-technology10To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.0095. ConclusionsThis study represents the first LCA on distributed recycling and provides a method to expand this classof LCAs beyond HDPE. The results of this LCA showed that distributed recycling of post-consumerHDPE for 3-D printing filament uses less embodied energy than the best-case scenario investigated fora high-population density city using centralized recycling. For centralized recycling in a low-densitypopulation case study involving substantial embodied energy use for transportation and collection thesesavings for distributed recycling were found to extend over 80%. These results have significantimplications for policy makers interested in reducing the energy and emissions associated with plasticconsumption and recycling. On the scale of U.S. yearly HDPE recycling this would amount to over100 million MJ of energy conservation and substantial GHG emissions reductions even in the best casescenario for centralized recycling. It seems clear that policies should be enacted to create incentives fordistributed recycling on environmental grounds. With the open-source 3-D printing network expandingrapidly the potential for widespread adoption of distributed recycling of HDPE represents a novel pathto a future of distributed manufacturing appropriate for both the developed and developing world withlower environmental impacts than the current system.AcknowledgmentsThe authors would like to acknowledge helpful discussions with D.R. Shonnard and G. Anzalone. Thisresearch was supported by Sustainable Futures Institute and the McArthur Research Internship atMichigan Tech.ReferencesAl-Salem, S.M., Lettieri, P., Baeyens, J., 2009. Recycling and recovery routes of plastic solid waste(PSW): A review. Waste Management 29, 2625–2643.American Chemistry Council and Association of Postconsumer Plastic Recyclers. 2010 United Statesnational post-consumer plastics bottle recycling report. American Chemistry Council; 2011.Arena U, Mastellone ML, Perugini F. Life Cycle assessment of a plastic packaging recycling system.The International Journal of Life Cycle Assessment 2003;8(2):92-8.Baechler C, DeVuono M, Pearce JM. Distributed recycling of waste polymer into RepRap feedstock.Rapid Prototyping Journal 2013;19:2.Björklund A. Recycling revisited - life cycle comparisons of global warming impact and total energyuse of waste management strategies. Resources, Conservation and Recycling 2005;44(4):309-17.Bolin CA, Smith S. Life cycle assessment of ACQ-treated lumber with comparison to wood plasticcomposite decking, J. of Cleaner Production 2011; 19(6–7): 620-629.11To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009Bow JK, Parvizi J. The History of Modern Total Knee Replacement. In: Parvizi J, Klatt B, editors.Essentials in Total Knee Arthroplasty. SLACK Incorporated; 2011.Braanker GB, Duwel JEP, Flohil JJ. Developing a plastics recycling add-on for the RepRap 3D-printer.Delft University of Technology; 2010. [cited 2012 Sept 6] Available from:http://reprapdelft.files.wordpress.com/2010/04/reprap-granule-extruder-tudelft1.pdfThe Cambridge - MIT Institute. The ImpEE Project: Improving Engineering Education. TheCambridge-MIT Institute; 2005. [cited 2012 Sept 1] Available from: http://www-g.eng.cam.ac.uk/impee/?section=topics&topic=RecyclePlasticsSugar Replaces Petroleum in Major Chemical Process. Chemical Engineering Progress2008;104(10):19-20.Craighill AL, Powell JC. Lifecycle assessment and economic evaluation of recycling: a case study.Resources, Conservation and Recycling 1994;17:75-96.Crane NB, Tuckerman J, Nielson GN. Self-assembly in additive manufacturing: opportunities andobstacles. Rapid Prototyping Journal 2011;17(3):211-7.Derraik JGB. The pollution of the marine environment by plastic debris: a review. Marine PollutionBulletin 2002;44:842-52.Electric Choice. Electricity Prices by State. Electric Choice; 2010. [cited 2012 Sept 5] Available from:http://www.electricchoice.com/electricity-prices-by-state.phpFeeley SR, Wijnen B, Pearce JM. Evaluation of Potential Fair Trade Standards for an Ethical 3-DPrinting Filament. (to be published, 2014).Franklin Associates. Life cycle inventory of 100% postconsumer HDPE and PET recycled resin frompostconsumer containers and packaging. Kansas: Prairie Village; 2011.Gebhardt A, Schmidt FM, Hotter JS, Sokalla W, Sokalla P. Additive manufacturing by selective lasermelting the realizer desktop machine and its application for the dental industry. Physics Procedia2010;5:543-9.Gibson I, Rosen DW, Stucker B. Additive manufacturing technologies: rapid prototyping to directdigital manufacturing. Physics Procedia 2010;5:543-54.Grant T, James K. Life Cycle Impact Data for Resource Recovery from C&I and C&D Waste inVictoria. RMIT University, Centre for Design 2005.12To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009Hammond G, Jones C. Inventory of carbon & energy (ICE) Version 1.6a. Bath U; 2008.Hischier, R., Weidema, B., Althaus, H. J., Bauer, C., Doka, G., Dones, R. Hischier, R., Hellweg, S.,Humbert, S., Kollner, T. Lorincik, Y., Margni, M. & Nemecek, T. Implementation of life cycle impactassessment methods. Ecoinvent report; 2010.Hood J. How green was my balance sheet. Policy Review 1995;74.Horton T. Waste management opens Detroit recycling center. Waste management; 2009. [cited 2012Sept 2] Available from: http://www.wm.com/about/press-room/2009/20090617_WM_Opens_Detroit_Recycling_Center.pdfKreiger M, Anzalone GC, Mulder ML, Glover A, Pearce JM. Distributed recycling of post-consumerplastic waste in rural areas. MRS Online Proceedings Library; 1492, mrsf12-1492-g04-06doi:10.1557/opl.2013.258 (2013).Kreiger M, Pearce JM. Environmental Impacts of Distributed Manufacturing from 3-D Printing ofpolymer components and products. MRS Online Proceedings Library; 1492, mrsf12-1492-g01-02doi:10.1557/opl.2013.319 (2013a).Kreiger M, Pearce JM. Environmental Life Cycle Analysis of Distributed 3-D Printing andConventional Manufacturing of Polymer Products, ACS Sustainable Chemistry & Engineering,Engineering; 2013b: 1 (12):1511–1519.Lebreton LC-M, Greer SD, Borrero JC. Numerical modeling of floating debris in the world’s oceans.Marine Pollution Bulletin; 2012:64(3):653-61.Lewis R, Sullivan JB. Toxic hazards of plastic manufacturing. Hazardous Materials Toxicology:Clinical Principles of Environmental Health. Baltimore: Williams & Wilkins; 1992. p. 505–15.Lotfi A. Plastic/Polymer Recycling; 1997. [cited 2012 Sept 1] Available from:http://www.lotfi.net/recycle/plastic.htmlLyman H. Lyman Filament Extruder. 2012. [cited 2012 Dec 2] Available from:http://www.thingiverse.com/thing:30642McCraw D, Motes JE. Use of plastic mulch and row covers in vegetable production. Division ofAgricultural Sciences and Natural Resources - Oklahoma State University. [cited 2012 Sept 1]Available from: http://pods.dasnr.okstate.edu/docushare/dsweb/Get/Document-1099/F-6034%20web.pdf13To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009McNaney T. Filabot Personal Filament Maker for 3D Printers. Filabot; 2012. [cited 2012 Aug 31]Available from: http://filabot.com/MiniRecyclebot – RepRapWiki. RepRapWiki; 2012. [cited 2012 Sept 12] Available from:http://reprap.org/wiki/MiniRecyclebotMissouri Department of Natural Resources. Recycling Economics: Higher Costs Are An Illusion. [cited2012 Sept 2] Available from: http://www.dnr.mo.gov/env/swmp/pubs-reports/rececon.htmOhio Department of Natural Resources (DNR) - Division of Recycling & Litter Prevention. PlasticsRecycling. [cited 2012 Sept 2] Available from: http://ohiodnr.com/tabid/17876/Default.aspxPachauri RK, Reisinger A. Climate change 2007: synthesis report. Contribution ofworking groups I, II and III to the fourth assessment report of the Intergovernmental Panel on ClimateChange. Geneva, Switzerland: Intergovernmental Panelon Climate Change; 2007.Pearce Corps. Global Hunger and Food Security Initiative. President’s global food security initiativefeed the future. PAPA with USAID/Senegal; 2011.Pearce JM, Blair C, Laciak KJ, Andrews R, Nosrat A, Zelenika-Zovko I. 3-D Printing of Open SourceAppropriate Technologies for Self-Directed Sustainable Development. J of Sustainable Development2010;3:17-29.Pearce JM. Building Research Equipment with Free, Open-Source Hardware. Science 2012;337(6100):1303–4.Pearce JM. Open-source lab: how to build your own hardware and reduce research costs. Elsevier(2014).Peças P, Ribeiro I, Folgado R, Henriques E, A Life Cycle Engineering model for technology selection:a case study on plastic injection moulds for low production volumes, J of Cleaner Production 2009;17(9) :846-856.Perugini F, Mastellone ML, Arena U. A Life Cycle Assessment of Mechanical and Feedstock RecyclingOptions for Management of Plastic Packaging Wastes. Environmental Progress 2005;24:137-54.Petrovic V, Gonzalez JVH, Ferrando OJ, Gordillo JD, Puchades JRB, Grinan LP. Additive layeredmanufacturing: sectors of industrial application shown through case studies. International Journal ofProduction Research 2010;49(4):1061-79.Powell JC. Lifecycle assessment and economic valuation of recycling. Journal of Environmental14To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009Planning and Management 1996;39:97-112.RecycleBot Team. RecycleBot. RecycleBot blog; 2010. [cited 2012 Sept 2] Available from:http://recyclebot.tumblr.com/archiveRedd A. Regionalization brings economies of scale to east coast. World Wastes 1993;36:40-5.Rees JF. The fate of carbon compounds in the landfill disposal of organic matter. Journal of ChemicalTechnology and Biotechnology 1980;30(1):161-75.Reich MC. Economic assessment of municipal waste management systems—case studies using acombination of life cycle assessment (LCA) and life cycle costing (LCC). Journal of CleanerProduction; 2005;13(3):253-63.Rosato D. Plastics Processing Data Handbook, 2nd ed. Springer – Verlag; 1997.Ross S, Evans D. The environmental effect of reusing and recycling a plastic-based packaging system.J of Cleaner Production 2003;11(5):561-71.Rydberg T. Cascade accounting in life cycle assessment applied to polymer recycling. PolymerRecycling (UK) 1995;1:233-41.Sandhill Plastic. Bottle recycling rate rises to 27.8 percent. Sandhill Plastic 2010. [cited 2012 Sept 2]Available from: http://www.sandhillplastics.com/news/sandhill-plastics-tidbits/recycled-bottles-increase-27Shen L, Haufe J, Patel MK. Product overview and market projection of emerging bio-based plastics.Copernicus Institute for Sustainable Development and Innovation-Utrecht University; 2009.Sells E, Smith Z, Bailard S, Bowyer A, Olliver V. RepRap: The replicating rapid prototyper:maximizing customizability by breeding the means of production. Robotica 2011;29(1):177-91.Song HS, Hyun JC. A study on the comparison of the various waste management scenarios for PETbottles using the life-cycle assessment (LCA) methodology. Resources, Conservation and Recycling1999; 27:267-84.Sparks D, Barnett ST. The informal sector in sub-saharan Africa: out of the shadows to fostersustainable employment and equity. International Business & Economics Research Journal 2010;9:1-11.Subramanian PM. Plastics recycling and waste management in the US. Conservation and Recycling.Resources, Conservation and Recycling 2000;28(3-4):253-63.15To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009Themelis, N. J., Castaldi, M. J., Bhatti, J., & Arsova, L. Energy and Economic Value of Non-recycledPlastics (NRP) and Municipal Solid Wastes (MSW) that are Currently Landfilled in the Fifty States.New York, NY: Columbia University, 2011.Thymark M. DIY shredder crusher chopper grinder. GrabCAD. [cited 2012 Sept 12] Available from:http://grabcad.com/library/diy-shredder-crusher-chopper-grinderTorcellini R. Plastic extruder for growing media. The Adventures of Bigelow Brook Farm blog. 2010.[cited 2012 Sept 2] Available from: http://web4deb.blogspot.com/2010/12/plastic-extruder-for-growing-media.htmlUpcraft S, Fletcher R. The rapid prototyping technologies. Assembly Automation 2003;23(4):318-30.U.S Energy Information Administration (EIA). State Energy Data System. EIA; 2010. [cited 2012 Sept12] Available from: http://www.eia.gov/state/seds/hf.jsp?incfile=sep_use/total/use_tot_TXa.html&mstate=TexasU.S. Energy Information Administration (EIA). FAQ-How much electricity does an American homeuse; 2011. [cited 2012 Sept 11] Available from: http://www.eia.gov/tools/faqs/faq.cfm?id=97&t=U.S. Environmental Protection Agency (EPA). Municipal solid waste in the United States: facts andfigures. EPA; 2010. [cited 2012 Sept 2] Available from:http://www.epa.gov/epawaste/nonhaz/municipal/msw99.htmU.S. National Park Service: Mote Marine Lab. Time it takes for garbage to decompose in theenvironment. [cited 2012 Sept 2] Available from:http://des.nh.gov/organization/divisions/water/wmb/coastal/trash/documents/marine_debris.pdfWittbrodt BT, Glover AG, Laureto J, Anzalone GC, Oppliger D, Irwin JL, Pearce JM. Life-cycleeconomic analysis of distributed manufacturing with open-source 3-D printers, Mechatronics 2013;23:713-726.Xue M. Research on polymer composites of replacement prostheses. International Journal ofBiomedical Engineering and Technology 2011;7(1):18-27.Zhang C, Anzalone NC, Faria RP, Pearce JM. Open-Source 3D-Printable Optics Equipment. PLoSONE 2013;8(3): e59840. doi:10.1371/journal.pone.005984016To be published: M.A. Kreiger, M.L. Mulder, A.G. Glover, J. M. Pearce, Life Cycle Analysis of Distributed Recycling ofPost-consumer High Density Polyethylene for 3-D Printing Filament, Journal of Cleaner Production, 2014. DOI;http://dx.doi.org/10.1016/j.jclepro.2014.02.009Figure CaptionsFigure 1. Schematic of Conventional HDPE Recycling.Figure 2. Schematic of Distributed HDPE Recycling.Figure 3: Map ofDetroit RecyclingCollection Centersand Great LakesRecycling Center.17