With the introduction of the self-replicating rapid prototyper (RepRap) 3-D printer [1
], the costs of additive manufacturing (AM) with 3-D printers have been reduced enough to be accessible to consumers. This has allowed for the emergence of a distributed manufacturing paradigm in AM [4
], where 3-D printing can be used to manufacture products for the consumer and by the consumer directly, with significant savings compared to the purchasing of mass-manufactured products [7
]. The emerging support for this model in the business literature [13
] is in part due to the exponential rise of free digital design file sharing for 3-D printed products [12
], which ranges from sophisticated scientific instruments [16
] to everyday toys for children and cosplayers [10
]. Regardless of the sophistication of the product, high return on investments (ROIs) can be enjoyed based on the download substitution values using commercial polymer 3-D printing filament [21
]. Commercial filament, however, is marked up significantly (e.g., >5×–10×) over the cost of the raw polymers, which limits the cost savings and thus, the deployment velocity of distributed manufacturing to further increase the rate of accessibility of AM [23
]. The negative effects of the high costs of filament are most notable for large format 3-D printers (those with a build volume that is greater than a cubic foot), which can process several kg of polymer in a single print lasting over 24 h.
One method of overcoming these cost barriers is to use a means of distributed plastic recycling, involving upcycling plastic waste into 3-D printing filament with a recyclebot (an open source waste plastic extruder [24
]). Previous research on the life cycle analysis of the recyclebot process found a 90% decrease in the embodied energy of the filament from the acquiring, processing of the natural resources, and the synthesizing compared to traditional filament manufacturing [25
]. This allows for the tightening of the loop of the circular economy [28
] because it enables real distributed recycling that eliminates nearly all energy use and pollution from transportation. Many recyclebot versions have been developed including open source variations from the Plastic Bank, Filastruder, Precious Plastic, Lyman, and Perpetual Plastic, as well as fully commercial versions including the Filastruder, Filafab, Noztek, Filabot, EWE, Extrusionbot, Filamaker (also has a shredder), and the Strooder, Felfil (OS) [29
]. Most recently, a “RepRapable Recyclebot” has been demonstrated [30
], where most of the machine’s parts can be 3-D printed from waste plastic themselves. Several polymers have been successfully recycled as single component thermoplastic filaments, such as polylactic acid (PLA) [30
], high-density polyethylene (HDPE) [24
], acrylonitrile butadiene styrene (ABS) [28
], elastomers [9
], as well as composites (e.g. waste wood [38
] and carbon fiber reinforced [39
Unfortunately, each time a polymer is heated and extruded (whether it be in the recyclebot filament making process or during conventional fused filament fabrication (FFF)/fused deposition modeling (FDM) 3-D printing), the mechanical properties are degraded [31
]. This effectively limits this process of recycling to 5 cycles [31
] without the use of blending virgin materials or adding materials for mechanical property reinforcement. To reduce the number of melt/extrude cycles of recycled plastic that is used for 3-D printing, one option is to eliminate the need for filament and print directly from pellets, flakes, regrind, or shreds of recycled plastic, which will be referred to as particles here. Several 3-D printers using fused particle fabrication (FPF) (or fused granular fabrication (FGF)) have been designed to accomplish this in the academic [42
], hobbyist [43
], and commercial systems [46
]. These systems have been tested with virgin pellets, however the mechanical properties of FPF printers using recycled polymer particles of various shapes and sizes has not been reported, which limits the ability of engineers to fabricate load bearing products from recycled waste using FPF 3-D printing.
In this study, the open source Gigabot X [51
], which is a large scale recycled plastic 3-D printer, was used to fabricate and test the mechanical properties of parts that were made using FPF to fill this knowledge gap. To establish a baseline, virgin PLA pellets were analyzed first and were then compared to four recycled polymers: (1) PLA regrind of 3-D printed parts (the most common 3-D printed plastic), (2) recycled ABS pellets (the second most common 3-D printed plastic), (3) recycled polyethylene terephthalate pellets (PET, the most common waste plastic [52
]), and (4) recycled polypropylene chips (PP, the second most common waste plastic [52
]). First, the material size characteristics were quantified using digital image processing. Then, a power and nozzle velocity matrix printing test were completed to determine the optimum print speed and temperature settings for a given polymer feedstock. During this phase of testing, any problems with bed adhesion and warping were identified and resolved. Third, a set of ASTM type 4 tensile bars were printed and were pulled to confirm the mechanical properties of the plastic when it was printed with a pellet drive extruder system and was compared with the results of past work with FFF/FDM 3-D printers, while noting the number of melt cycles.
The advantages of printing with recycled particles rather than filament include lower costs, because of the lower cost of the starting material, and it is also easier to print large objects (e.g., where more than 1 spool of material is required) [78
]. This has caused a recent surge in research that is related to 3-D printing with pellets including new printer designs: using a double stage screw [79
], derivative of a metal injection molder [80
], those based on RepRap technology [82
], and industrial robots for 3-D printing [83
], as well as a large range of polymers including conventional filament materials and recycled materials [82
] as well as conductive polymer composites [84
] and flexible materials [85
]. These studies all indicate that FPF 3-D printing will play a larger role in the future of the 3-D printing industry and are corroborated by the results of this study.
The Gigabot X successfully printed with a wide range of particle sizes and distributions, which opens up a large array of starting materials beyond high-quality (e.g., uniform size and spherical) pellets. It can be concluded that the Gigabot X and similar FPF printers with good auger-barrel tolerances (+/−0.025 mm) can handle a wide range of polymer inputs, including recycled materials with minimal post processing (i.e., only cleaning and grinding/shredding). In addition, due to the ease of extrusion and the ability to print in a wide distribution of particle sizes, this system is an ideal candidate for upcycling waste plastics and the development of unique plastics co-polymers and composites. The mechanical testing using tensile strength indicated that FPF did not degrade the polymer properties (e.g., similar to FFF), however, future work should also consider testing other mechanical properties such as compression, impact resistance, fracture toughness, creep testing, fatigue testing, and flexural strength. This will result in additional potential applications of recycled polymer FPF.
The ability to define multiple heating zones in the FPF extrusion system as in the Gigabot X is useful for multiple reasons. First, by establishing a pre-heating zone prior to extrusion, this helps to both transfer the required thermal energy to the plastic (given the high material flow rate) and to achieve sufficiently low viscosities for printing. The low viscosity also reduces motor wear from over torqueing and allows for faster material throughput during purging cycles. Another benefit of having multiple heating zones is to allow for more custom temperature profiles. Some plastics experience more shear heating than others. Therefore, having a descending temperature gradient would allow the temperature at the inlet to be higher to start melting and mixing the pellets as soon as possible to allow for the best throughput, while simultaneously setting the die heater at a lower temperature to prevent degradation from overheating in the shearing (metering) section. Changing the screw geometry or the compression ratio for every plastic would be very unsustainable, so being able to have control over how the different plastics flow while they are inside the extruder without having to replace the screw, comes down to controlling speed and temperature profiles.
The large format FPF type printer has unique abilities to print large components at high mass flow rates, resulting in dramatically shorter print times for large components. This is due to the ability to easily print with large nozzle sizes (1–2 mm). This is particularly useful in the manufacturing of large, functional components with non-critical surface smoothness criterion. Some applications include custom furniture, complete sporting goods equipment, large research tools, breathable casts and other health care products, agricultural processing equipment, OEM components, construction applications, and on demand fixtures or jigs.
There are, however, some limitations with this prototype system, including lower than normal FFF resolution in the x-y-plane (1.75 mm diameter minimum due to the nozzle diameter) and further limitations are imposed on the component size due to the high heat transfer rates from the large (comparative to traditional FFF printers) contact area of the printer’s hot-end, meaning that parts that are less than 20 mm by 20 mm cannot be printed reliably. Useful future work would consider other mechanisms for providing FPF beyond the system that has been described here. The Gigabot X is also currently lacking in any sort of part cooling system, which is a common feature on most FFF printers to assist in the cooling of the extruded plastics. This allows the printer to print complex geometries, such as overhangs of plastic, without losing dimensional accuracy or creating visual blemishes on the surface of the print. Without this, Gigabot X is limited in the geometries that it is capable of printing accurately. Finally, the prototype provides approximately six hours of continuous feed before manual replenishment. If this technology is to be scaled commercially, users would require much longer printing sessions and an automated feeding system.
Future work is needed to quantify the environmental benefits of using FPF over conventional FFF/FDM with both conventional filament as well as recycled filament. Several plastic pellet types and sizes from multiple vendors were used in this research. Significant variability in the size and the uniformity of the sourced pellets was observed between the different vendors. Further research on this hardware is needed to support supplier variability using non-uniform flake and/or pellets. Finally, a detailed cost analysis is needed to quantify the economic benefits of utilizing this approach.
The Gigabot X system presents a uniquely robust open source solution to large format FPF printing. Notable with this design was the ability to print a large variety of polymers at dramatically lower print times when compared to traditional FFF type printing. As recycled plastic and pellets are less costly than filament, this system succeeds in lowering the economic barriers to the fabrication of large format, high value, plastic components, which has been an unfulfilled gap in the open source, distributed manufacturing design space.
The tensile strengths of the printed parts for traditional FFF/FDM prints and the FPF prints are comparable for all of the polymers that were tested when the melt/print cycles are taken into account. This means that there will be no need to sacrifice part strength by using FPF systems. More studies need to be conducted to determine how the layer adhesion compares between an FFF and an FPF system.
Finally, a novel methodology was tested to successfully operate any FPF system (and the Gigabot X in particular) to experimentally determine the best 3-D printing parameters for new polymers. The line test was developed to show the best temperatures to print the new polymer regardless of type or particle size distribution. The single walled vase test was used to determine the theoretical vs actual mass and to get a flow percentage that would calibrate the extruder. The vase test was also used to determine the actual extrusion width to be put into Slic3r to prevent overlapping lines or under-extrusion. These simple tests can be performed over the course of a couple days for any unknown polymer and, once completed, should give the correct settings for successful prints.