Original Article

Substrate Release Mechanisms for Gas Metal Arc Weld 3D Aluminum Metal Printing

Amberlee S. Haselhuhn ,1 Eli J. Gooding , 1 Alexandra G. Glover , 1 Gerald C. Anzalone , 1 Bas Wijnen , 1 Paul G. Sanders , 1 and Joshua M. Pearce 1,2

Abstract

Limited material options, prohibitively expensive equipment, and high production costs currently limit the ability of small and medium enterprises to use 3D printing to prototype and manufacture metallic goods. A low - cost open - source 3D metal printer that utilizes gas metal arc welding technology has been developed that could make metal printing accessible to the average consumer. Unfortunately, this technology would demand access to expensive cutting tools for part removal from the substrate. This article investigates several substrate treatments to provide a low- cost method to easily remove 3D- printed 1100 aluminum parts from a reusable substrate. Coatings of aluminum oxide and boron nitride on 1100 aluminum and A36 low - carbon steel substrates were tested. Lap shear tests were performed to assess the interlayer adhesion between the printed metal part and the print substrate. No warping of the substrate was observed during printing. It was determined that boron nitride- coated low- carbon steel provided the lowest adhesion strength. Printing aluminum on uncoated low- carbon steel also allowed easy removal of the aluminum part with the benefit of no additional coating steps or costs.

Introduction 3D metal printing is commercially electron beam freeform fabrication,11 and available in several forms: laser - based microwelding12,13 in a single - layer 3 D printing has the potential to additive manufacturing, weld- based multipass welding regime. Parts revolutionize the way goods are additive manufacturing, and shape produced by weld- based additive manufactured on a global scale, enabling deposition modeling. Laser - based manufacturing are inexpensive and mass customization while reducing both additive manufacturing methods include nonporous with good interlayer capital investment and production powder bed fusion (direct metal laser adhesion, but have a limited print costs.1,2,3,4 The ability of small and sintering), 5 selective laser sintering, 6 resolution and poor surface finish. medium enterprises (SMEs) to make selective laser melting, 7 and directed Microwelding is the exception, exhibiting full use of this potential is currently energy deposition (laser cladding). 6 excellent dimensional control and finer restricted by material choice. Most These methods offer excellent surface finish resulting from the small- SMEs only have local access to polymeric dimensional control but have large diameter electrode and wire employed. fused filament fabrication (plastic production costs due to the use of lasers material extrusion). Third - party or metal powders. Shape deposition manufacturing prototyping services offer a broader processes feature both additive and range of print media, including Weld - based additive manufacturing subtractive manufacturing.14,15,16 A single ceramics and metals. Most commercial methods include gas metal arc welding layer of metal is melted, sintered, or 3D metal printers are out of reach of (GMAW), 8,9 gas tungsten arc welding, 10 welded in a rough net shape and then SMEs as they typically cost more than directed energy deposition and powder subsequently milled down to a precise $ 500,000. bed fusion (electron beam melting), 7 geometry before the next layer of metal is

Departments of 1 Materials Science & Engineering and 2 Electrical & Computer Engineering, Michigan Technological University , Houghton, Michigan.

204 3D PRINTING MARY ANN LIEBERT, INC. • VOL. 1 NO. 4 • 2014 • DOI: 10.1089/3dp.2014.0015 Substrate Release Mechanism

printed. These processes have excellent dimensional control but are expensive and time - consuming as both printing and computer numerical control (CNC) milling equipment are required to produce a part.

The high cost and slow throughput of the current commercial metal 3D printers limit their application to expensive finished products such as custom hip replacements and maxillofacial repairs.17 There is thus an urgent need for a low- cost 3D printer capable of depositing metals directly for both rapid prototyping and rapid manufacturing for SMEs.

Recent development of a low- cost open- source 3D metal printer exploiting GMAW technology offers the potential for the general public to 3D print metal parts.18 This printer, based on the RepRap concept19 (self- rep licating rap id prototyper), is capable of being partially manufactured either by itself or produced in an incremental fashion from components produced by a standard fused filament fabrication system. It utilizes a moving stage upon Figure 1. Labeled photograph of the three - axis stage with attached aluminum weld gun. which the substrate is placed and a fixed print head comprised of a workshop GMAW. Aluminum and steel parts have interface between the 3D- printed part metal arc welder and a three- axis stage as been printed with this printer using and substrate was quantified using a lap shown in Figure 1 . 16 The GMAW, a single - pass, multilayer welding and were shear measurement commonly employed Millermatic 140 with Spoolmate 100 removed from the substrate with a to assess the strength of adhesives.20 The weld gun, supplied the material used to vertical band saw. This method of part results are discussed and a generalized print and the energy required to melt the removal is suboptimal as it requires mechanism is proposed for enabling material. The three - axis stage was additional processing equipment, costs, substrate selection for aluminum microprocessor controlled, permitting and time. Thus, an alternative method is GMAW 3D printing. precise CNC of both the position and desired. speed of the platform upon which parts were printed. The part was built upon a This article investigates several substrate Materials and Methods sacrificial 6.35- mm - thick mild steel or treatments to provide low - cost release aluminum plate. mechanisms to remove 3D - printed Description of a 3D Metal Printer aluminum metal parts from the print The stage was derived from an open - substrate, outlined in Table 1. The The metal printer comprised two distinct source 3D printer design known as a substrate adhesion and strength of the components, a workshop- grade gas Rostock, 21 which is a RepRap derivative.

Table 1. Substrate release mechanisms analyzed by this study

Mechanism Print material Coating Coating thickness (µ m) Substrate

1 ER1100 aluminum None 0 1100 aluminum 3 ER1100 aluminum Aluminum oxide 18.8 1100 aluminum 3 ER1100 aluminum Boron nitride 5.9 1100 aluminum 2 ER1100 aluminum None 0 A36 low - carbon steel 3 ER1100 aluminum Aluminum oxide 18.8 A36 low - carbon steel 3 ER1100 aluminum Boron nitride 5.9 A36 low - carbon steel

MARY ANN LIEBERT, INC. • VOL. 1 NO. 4 • 2014 • DOI: 10.1089/3dp.2014.0015 3D PRINTING 205 Haselhuhn et al.

The original Rostock printer had the source CAD package, was used to develop substrate at a time, held at an angle until extruder mounted on the moving end the solid models, which were then sliced the substrate was completely covered. effector, whereas the three- axis stage with the 3D printing software Cura 25 and The slurry was remixed between used in this work was essentially a converted into G - code. G - code provided substrate coatings to minimize settling. Rostock turned upside down, with the numerical control to the stepper motors, The coated substrates were laid flat to workpiece on the moving end effector directing them when to move and how dry for approximately 30 minutes. and the “ extruder ” (welding gun) fixed in fast to move. The metal printer interfaced position above it. with these programs using a printer An aerosol - based boron nitride coating server developed at Michigan Tech with a (ZYP Coatings), approximately 5.9 μ m During this study, welding parameters web - based interface. 26 thick, was also sprayed directly onto the were set manually and the motion of the two substrate types. The coating stage was adjusted to produce a quality Materials Preparation and Printing thicknesses of aluminum oxide (18.8 μm) bead. A quality bead was defined as a of Test Specimens and boron nitride (5.9 μ m) were chosen continuous line of 3D- printed material to prevent conductivity issues between with consistent profile. Argon shield gas Standard ER1100 aluminum GMAW the welder and substrate that might was used to minimize inclusions and wire, 0.030 inches (0.762 mm) in prevent arc formation and thus preclude spatter so as to produce a higher quality diameter, was used as the weld filament, welding. Four samples per treatment weld bead. Flux core wire was not utilized while degreased 1100 aluminum and condition were prepared. as it can leave a waste layer on top of the A36 low- carbon steel were used for the weld, making it difficult to print multiple print substrate materials. Three substrate Lap shear test samples were prepared layers. release mechanisms were tested, with aluminum substrate coupons incorporating the use of various coatings (50.8 × 76.2 × 6.35 mm 3 ) butted together Details of the Three - Axis Stage and substrate materials to modify with a 25.4 × 50.8 × 6.35 mm3 print made adhesion of the 3D - printed part (Table 1). The three - axis stage is shown in Figure 1. over the butt joint ( Figure 2 ). A sheet of An aluminum oxide coating, All of the designs for the hardware and 6.35- mm - thick A36 low- carbon steel approximately 18.8 μ m thick, was applied all of the software employed are free was placed under the substrate plates to to the two substrate types. Coating and open - source. 22 The all - metal act as a heat sink. Relevant print thickness was calculated using an average construction 23 minimized risk of damage parameters, including cover gas, wire mass across the surface area of the feed rate, power, and slicing speed are

due to weld spatter and heat. The drive substrate. The coating was applied as a mechanism utilized three NEMA17 shown in Table 2. The slicing speed is slurry prepared by mixing 100 mL 95 % stepper motors (5.5 kg · cm torque) with defined as the speed at which the three- isopropyl alcohol (BDH, Inc.) with 27 g lead screws integrated into their shafts, axis stage, upon which the substrate of 10 μ m aluminum oxide powder of requiring no couplings between the rests, moves. Each test specimen was 99.7% purity (Sigma- Aldrich). Mixing motors and lead screws. The trapezoidal - water - quenched immediately following was performed with a magnetic stir bar threaded lead screws had a 5 mm pitch print completion. An identical procedure for 5 minutes to ensure homogeneity. and were 300 mm in length. The three was maintained for printing on A36 low- The slurry was poured onto a single motors were arranged vertically on a carbon steel coupons. 394 mm circle, spaced 120° apart as shown in Figure 1. In general, the three- axis stage was based upon an industrial delta robot design commonly used for pick- and - place operations, except allowing for greater movement in the z - direction. Control was provided by an Arduino - based controller. Firmware (software resident on the printer ’ s microcontroller) Figure 2. Schematic drawing of the 3D - printed aluminum lap shear test specimen. controlled the motion of the printer, translating commands from a printer server running on a host computer. The host computer, in turn, served a web Table 2. Print parameters for 3D printing ER1100 aluminum interface from which the end user was Weld unit Wire feed Slicing able to control stage motion, queue print Substrate voltage rate speed jobs, and make configuration changes. material setting Power (W) (in./ min) (mm/ min) Cover gas

Software Tool Chain ER1100 3 854 90 180 aluminum Argon RepRap 3D printers utilize A36 low - 1 764 75 stereolithography (.stl) files for the input. carbon steel OpenSCAD,24 a script- based open-

206 3D PRINTING MARY ANN LIEBERT, INC. • VOL. 1 NO. 4 • 2014 • DOI: 10.1089/3dp.2014.0015 Substrate Release Mechanism

Testing and Analysis of Samples model indicated that there was a aluminum.27 Chiefly, the compound significant difference in the breaking Fe 2Al 5 forms adjacent to the ferrous The adhesion strength of the interface strength between the samples printed on substrate and FeAl3 forms adjacent to the between the 3D- printed part and the low- carbon steel versus aluminum. aluminum.28,29 This formation is substrate was quantified using a lap shear There was also a significant difference in dominated by the diffusion of iron atoms test. The substrate materials were loaded the breaking strength between samples into the liquid aluminum26 and is until failure by an MTS tensile load printed on boron nitride compared with affected by factors such as the presence frame with a 150 kN load cell at a rate of aluminum oxide or no coating. Two - of silicon in the aluminum30 and carbon 85 μ m per second. Failure was marked as sample t - tests (α = 0.05) confirmed these in the iron, 31 temperature, and interaction the maximum stress required to break results. No deformation of the print time between the aluminum and iron the printed metal coupon away from one substrate was observed. atoms. 32 While most manufacturers try of the substrate plates. The maximum to suppress the formation of aluminum – load at failure corresponded to the The first substrate release mechanism, a iron intermetallic compounds due to strength of the interface for a given control group, examined the adhesion their brittle nature, here the formation of adhesion area. strength between commercially pure these intermetallics was encouraged to aluminum printed on commercially pure form a brittle interface between the aluminum. In this instance, good joining printed part and substrate. This brittle Results and Discussion between the printed and substrate interface may have allowed the materials was observed. This result was aluminum samples to be easily removed The results of the lap shear test are expected because no compounds or from the low - carbon steel substrate, not summarized in Figure 3. The error bars coatings were applied to prevent only via a lap shear test but also with a in this figure represent ± 1 standard error adhesion. hammer and chisel, exploiting the low of the mean, which is the standard strength of this interface. deviation normalized for sample size. The second substrate release mechanism Overall, it was observed that the peak attempted to exploit the formation of Printing aluminum on uncoated steel breaking stress of aluminum printed on intermetallic phases between aluminum worked well to prevent adhesion between low- carbon steel was less than that of and low- carbon steel. Intermetallic the sample and the substrate. This aluminum printed on aluminum, phases form between solid solutions of behavior was likely the result of two regardless of coating type. In all samples, different metals and are characterized as factors: the formation of aluminum – the peak breaking stress of boron nitride - having a crystal structure that is different iron intermetallic compounds and coated substrates was less than both the from each individual solid solution minimal weld penetration into the aluminum oxide - coated substrates and phase. Aluminum– iron intermetallics steel due to differences in thermal the uncoated substrates. A two - way have been well studied, perhaps because properties between aluminum and steel. analysis of variance (ANOVA; α = 0.05) iron is a major impurity element in However, future studies encompassing compositional analysis and microstructural characterization are recommended to test these hypotheses.

Both the specific heat capacity and melting temperature of aluminum are different from those of steel. For instance, the specific heat capacity of 1100 aluminum is 0.90 J/ (g · K), whereas the specific heat capacity of A36 low- carbon steel is 0.48 J / (g · K). This indicates that more energy is required to increase the temperature of the aluminum as compared to steel. Generally, a mass of aluminum will diffuse heat more quickly than the same mass of steel due to aluminum ’ s higher thermal conductivity. Aluminum also has a higher heat capacity allowing it to store more thermal energy before reaching temperatures yielding weld penetration in steel. These two factors allow aluminum to store more thermal energy than steel at a given Figure 3. Results of the lap shear test are shown for ER1100 aluminum printed on both 1100 temperature. This heat storage acts as an aluminum and A36 low - carbon steel plate, including substrates coated with aluminum oxide impetus to allow surface chemical (alumina) and boron nitride (BN). Error bars represent ± 1 standard error. reactions, such as the formation of

MARY ANN LIEBERT, INC. • VOL. 1 NO. 4 • 2014 • DOI: 10.1089/3dp.2014.0015 3D PRINTING 207 Haselhuhn et al.

intermetallic compounds, to occur. As both an economic and environmental aluminum on uncoated low - carbon steel, described previously, a thin intermetallic benefit. Preliminary work indicated that additional work is necessary to establish layer can form a very low - strength, the same substrate can be used several the effect of aluminum alloy on wetting brittle interface between the aluminum times. The aluminum oxide and boron and adhesion to the substrate. and steel substrate that allows the nitride coatings can be scraped, sanded, aluminum to be removed from the steel or simply washed off with water to with ease. As the steel remains solid at prepare the surface for reuse, an Conclusions the interface due to differences in melting improvement over other 3D printing temperatures, there is much less mixing techniques such as laser melting and This study has provided low- cost and diffusion of the two metals, laser welding of metals, which require a methods allowing easy removal of 3D - preventing both adhesion and warping sacrificial substrate. printed aluminum parts from readily of the substrate during printing. One available substrates. Printing aluminum additional benefit of this substrate release This work enhances the value of the low - on boron nitride- coated low- carbon steel mechanism is that it does not require cost open- source 3D metal printer by substrates produced the weakest additional coating materials or processes providing a simple means for removing interfacial adhesion strength. Aluminum as described by the other proposed parts from the substrate and by parts printed on uncoated low - carbon mechanisms. While the ANOVA and permitting substrate reuse. Prototyping steel could also be easily removed with two- sample t - tests do not indicate that and manufacturing printed metal parts the benefit of not requiring application of this mechanism provides the lowest- using this technology is not limited to a coating prior to printing. strength interface, it is a low- cost, no- just SMEs in developed regions, but also effort method to prevent adhesion enables metal 3D printing of open- between the 3D- printed aluminum part source appropriate technology 33 in the Acknowledgments and substrate. developing world.34 Clearly, the ability to produce custom functional metal parts The authors would like to acknowledge The third substrate release mechanism, (e.g., bicycle components, water pump P. Fraley and J. Kolacz for technical sacrificial aluminum oxide and boron components, or small wind turbines) in assistance, and valuable discussions with nitride coatings, is frequently used in the a relatively isolated community would R. Gorham and America Makes. This metal casting industry to prevent have far- reaching implications. Beyond material is based on research sponsored adhesion of liquid aluminum metal to the economic benefits, this technology by Air Force Research Laboratory under iron - based permanent molds, providing may also have utility in education as Agreement Number FA8650- 12 - 2 - 7230. the motivation for assessing the efficacy assessed in work by UNESCO of similar treatments on 3D metal considering how 3D printing could be printing substrates. Application of boron used for education in such communities. 35 Disclaimer nitride coatings to a print substrate limits adhesion between 3D - printed metallic Evaluation of polymer - printing RepRaps The U.S. Government is authorized to parts and metallic substrates, allowing demonstrated significant cost savings3 reproduce and distribute reprints for the parts to be removed with relative and environmental benefits36 as governmental purposes notwithstanding ease. Even at more than triple the coating compared to conventional manufacturing any copyright notation thereon. The thickness, aluminum oxide (18.8 μm methods; a similar evaluation is needed views and conclusions contained herein coating) exhibited a higher adhesion to determine the economics of this metal are those of the authors and should not strength than did boron nitride (5.9 μm 3D printing method. Future work is also be interpreted as necessarily representing coating). In fact, there was no statistical required to determine substrate the official policies or endorsements, difference between the adhesion strength durability and life. This will establish a either expressed or implied, of Air Force of noncoated substrates versus aluminum reasonable substrate reuse rate as well as Research Laboratory or the U.S. oxide- coated substrates. This behavior the effect on life cycle economics. Government. may be due to the fact that the aluminum Characterization of the printed metal oxide coating did not contain any part is necessary to establish chemical binders, whereas the boron compositional and mechanical Author Disclosure Statement nitride coating did. These chemical properties and their relationship to the binders may help the coating survive the type of coating used, weld parameters, No competing financial interests exist. harsh welding conditions. Some of the and local environment. This may be aluminum oxide coating may have been accomplished by compositional removed by the energy associated with mapping, hardness measurements, and References rapid heating of the substrate and shield standard tensile and bend tests. The gas, limiting its ability to prevent part– substrate interface also requires 1. Markillie P. A third industrial adhesion between the molten aluminum similar characterization to better define revolution: Special report: Manufacturing and metallic substrate. the mechanisms responsible for adhesion and innovation. The Economist April 21, modification and to provide insight for 2012. All of the techniques outlined in this additional manipulation of those 2. Gershenfeld N. Fab: The coming article allow for substrate reuse, which is mechanisms. In the case of printing revolution on your desktop— from

208 3D PRINTING MARY ANN LIEBERT, INC. • VOL. 1 NO. 4 • 2014 • DOI: 10.1089/3dp.2014.0015 Substrate Release Mechanism

personal computers to personal intelligent art of direct metal rapid 28. Bouch é K, Barbier F, Coulet A. fabrication. Basic Books, New York, NY, tooling process. Robot Comput Integr Intermetallic compound layer growth 2005. Manuf 2006;22:113– 123. between solid iron and molten 3. Wittbrodt BT, Glover AG, Laureto J, 15. Song Y - A, Park S, Choi D, et al. 3D aluminum. Mater Sci Eng A Struct Mater et al. Life- cycle economic analysis of welding and milling: Part I — a direct 1998;249:167– 175. distributed manufacturing with open- approach for freeform fabrication of 29. Kobayashi S, Yakou T. Control of source 3 - D printers. Mechatronics metallic prototypes. Int J Mach Tool intermetallic compound layers at 2013;23:713– 726. Manuf 2005;45:1057– 1062. interface between steel and aluminum 4. Lipson H, Kurman M. Fabricated: The 16. Song Y- A, Park S, Chae S- W. 3D by diffusion- treatment. Mater Sci Eng A new world of 3D printing. John Wiley & welding and milling: Part II— optimization Struct Mater 2002;338:44– 53. Sons, New York, NY, 2013. of the 3D welding process using an 30. Akdeniz MV, Mekhrabov AO, Yilmaz 5. Khaing MW, Fuh JYH, Lu L. Direct experimental design approach. Int J T. The role of Si addition on the interfacial metal laser sintering for rapid tooling: Mach Tool Manuf 2005;45:1063– 1069. interaction in Fe- Al diffusion layer. Scr Processing and characterisation of EOS 17. McGurk M, Amis AA, Potamianos P, Metall Mater 1994;31:1723 – 1728. parts. J Mater Process Technol et al. Rapid prototyping techniques for 31. Shih TS, Tu SH. Interaction of steel 2001;113:269– 272. anatomical modelling in medicine. Ann with pure Al, Al - 7Si and A356 alloys. 6. Costa Santos E, Shiomi M, Osakada K, R Coll Surg Engl 1997;79:169 – 174. Mater Sci Eng A Struct Mater 2007;454 – et al. Rapid manufacturing of metal 18. Anzalone GC, Zhang C, Wijnen B, 455:349 – 356. components by laser forming. Int J Mach et al. A low- cost open- source metal 3- D 32. Borrisutthekul R, Yachi T, Miyashita Tool Manu 2006;46:1459– 1468. printer. IEEE 2013;1:803– 810. Y, et al. Suppression of intermetallic 7. Murr LE, Gaytan SM, Ramirez DA, 19. Jones R, Haufe P, Sells E. RepRap — reaction layer formation by controlling et al. Metal fabrication by additive the replicating rapid prototyper. Robotica heat flow in dissimilar joining of steel manufacturing using laser and electron 29;2011:177– 191. and aluminum alloy. Mater Sci Eng A beam melting technologies. J Mater Sci 20. ASTM Standard D1002, 2010, Struct Mater 2007;467:108– 113. Tech 2012;28:1– 14. Standard test method for apparent shear 33. Pearce JM. The case for open source 8. Zhang YM, Li P, Chen Y, et al. strength of single- lap - joint adhesively appropriate technology. Env Dev Sust Automated system for welding - based bonded metal specimens by tension 2012;14:425– 431. rapid prototyping. Mechatronics loading (metal - to - metal). ASTM 34. Pearce JM, Morris Blair C, Laciak KJ, 2002;12:37– 53. International, West Conshohocken, PA, et al. 3- D printing of open source 9. Zhao H, Zhang G, Yin Z, et al. A 3D 2014. DOI: 10.1520/ D1002 - 10. appropriate technologies for self - directed dynamic analysis of thermal behavior 21. Reprap Org. Rostock. http:/ / reprap sustainable development. J Sust Dev during single - pass multi - layer weld - . org / wiki / Rostock (last accessed June 24, 2010;3:17– 29. based rapid prototyping. J Mater Process 2014). 35. Canessa E, Fonda C, Zennaro M. Technol 2011;211:488 – 495. 22. Michigan Tech’ s Open Sustainability Low- cost 3D printing: For science, 10. Wang H, Jiang W, Ouyang J, et al. Technology Lab. Open - Source Metal education & sustainable development. Rapid prototyping of 4043 Al- alloy parts 3 - D Printer. www.appropedia.org / Open - ICTP, Trieste, Italy, 2013. by VP - GTAW. J Mater Process Technol source_metal_3 - D_printer (last accessed 36. Kreiger M, Pearce JM. Environmental 2004;148:93– 102. June 24, 2014). life cycle analysis of distributed 3 - D 11. Wanjara P, Brochu M, Jahazi M. 23. Michigan Tech’ s Open Sustainability printing and conventional manufacturing Electron beam freeforming of stainless Technology Lab. Open - Source Metal of polymer products. ACS Sust Chem steel using solid wire feed. Mater Des 3 - D Printer v2. www.appropedia Eng 2013;1:1511– 1519. 2007;28:2278– 2286. .org / MOST_open - source_metal_3 - D_ 12. Katou M, Oh J, Miyamoto Y, et al. printer_v2 (last accessed June 24, Freeform fabrication of titanium metal 2014). Address correspondence to: and intermetallic alloys by three - 24. OpenSCAD. www.openscad.org (last Joshua M. Pearce dimensional micro welding. Mater Des accessed June 24, 2014). Department of Materials Science & 2007;28:2093– 2098. 25. Ultimaker. Software Downloads. Engineering 13. Horii T, Kirihara S, Miyamoto Y. http: / / software.ultimaker.com / (last Michigan Technological University Freeform fabrication of superalloy accessed June 24, 2014). 601 M&M Building objects by 3D micro welding. Mater Des 26. https://github.com/mtu-most/franklin. 1400 Townsend Drive 2009;30:1093– 1097. 27. Belov NA, Aksenov AA, Eskin DG. Houghton, MI 49931 - 1295 14. Akula S, Karanukaran KP. Hybrid Iron in aluminum alloys. Taylor & adaptive layer manufacturing: An Francis, New York, NY, 2002. E- mail: pearce @ mtu.edu

MARY ANN LIEBERT, INC. • VOL. 1 NO. 4 • 2014 • DOI: 10.1089/3dp.2014.0015 3D PRINTING 209