Additive manufacturing can be construed quite broadly, but the core elements include controlled fusing of material(s) under computer control. Beginning in the 1980s, the field has expanded into many materials, being fused with a variety of techniques, resulting in builds with many different characteristics. From the earliest days, 3D printing as a field has evolved in many directions:
Even given innovations in the 2014–2018 period, much remains constant. Three basic technologies dominate the general market: toothpaste-tube-like extrusion, vat-based solidification of light-sensitive polymers, and powder-bed methods using adhesive binders, heat, and light or electron beams to shape plastics and metals. Lasers can be involved, but are not necessarily required. Heat is often a component of the process, either before fusing, at the moment of melting or near-melting, and in ovens after building for various types of heat-treating. Regardless of the medium—powder, filament, wire, gel, liquids, aerosols, cells, concrete—the constant factor is computer control of the deposition process. 3D printing is essentially a branch of robotics.
As of 2012, the ISO/ASTM standard 52900 specified the terminology for additive manufacturing. Even as this apparent standardization was underway, however, printer manufacturers continued to blur the lines between additive and subtractive technologies (as a new generation of hybrid machines came to market) and between the categories of addition. The most visible example of the latter boundary-crossing is HP’s Multi Jet Fusion technology, which touches on as many as four of the canonical categories.1 The ISO/ASTM model still has utility, however. Seven primary additive technologies were named, some of which have manufacturer-specific variants.2 They are as follows:
SLA was a pioneering technology, patented in 1986 by Charles Hull and brought to market in 1988 by his company, 3D Systems, which remains a leader in the field. It was the first of a class of additive technologies to build parts in a liquid medium of photo-curable resin using light (often in the UV spectrum, sometimes from a laser), with successive thin layers of material being added on top of each other. The primary advantages of SLA are 1) that it can be run unattended, 2) the wide range of build volumes, 3) the accuracy of the build (it is widely used in dental applications) combined with 4) the high quality of surface finish, and 5) the wide range of available materials. SLA has the downsides of requiring post-curing in some cases, and requiring post-processing that adds time and tediousness, especially when such work has the potential to damage the build. SLA is difficult to adapt for use as a multi-material method. Photo-curable resins can also be expensive compared to other alternatives.
While SLA is well established, it has recently been joined by a new innovation using the same basic principles. CDLP (also known as CLIP: continuous liquid interface production at Carbon, a young company launched in 2014 that pioneered the technology) behaves similarly to SLA: light interacts with a liquid polymer layer by layer. As in some SLA printers, direct light processing flashes an entire layer at once rather than tracing the shape path-wise with a laser. CLIP enjoys a speed advantage related to its bottom-up printing path combined with a proprietary light- and oxygen-permeable window at the bottom of the vat. This innovation removes the necessity of repeatedly separating freshly built layers from the bottom window, as in other SLA bottom-up printers.3 Many different kinds of resins can be used. Advantages include high speed relative to other additive methods, fine details, and a smooth surface finish. Disadvantages include brittleness in some produced builds, and uncertainty with regard to Carbon’s business model of not transferring ownership of the printers to customers.
Another variation of light-cured polymers is used in what is known as 2 photon polymerization. It can generate nano-scale features with layer thickness as small as 0.1 micron. The resulting structures (including cell scaffolds, as we will see in chapter 7) are used in optical, life-science, and other highly precise environments. The German firm Nanoscribe is an early leader in the emerging field.
| Vat Photopolymerization | |
|---|---|
|
Method of binding |
Light |
|
Selected materials |
Photopolymer resin, ceramics |
|
Layer thickness |
.025–.1 mm |
|
Support structures required? |
Yes |
|
Selected vendors |
3D Systems, Carbon, Formlabs, EnvisionTEC |
|
Advantages |
Smooth surface, high resolution, machine reliability, large build envelopes |
|
Limitations |
Photopolymers lack the strength of injection molded parts and degrade in sunlight |
Introduced just after SLA, FDM was developed in the late 1980s and brought to market in 1992. A leading vendor of these machines is Stratasys of the United States. It uses an extrusion process fed by a filament of nylon, ABS, or other plastics to build both prototypes to test for fit, feel, and other attributes, as well as production products. Certain FFF machines use two print heads, one with the primary structure’s material and another with a release material that is used to build support structures that are removed upon completion of the overall piece. Material flowing out of the print head assumes an oval profile as gravity takes hold, meaning that there are small but visible valleys between layers of material.
Heat is applied to the filament at the print head, and the resulting semiliquid state allows the material to be deposited in thin layers. FFF is able to create production parts 85 percent as strong as molded pieces, can build overhangs with integral but temporary support, supports multi-material builds with simple filament changes, and features relatively large build envelopes at the high end of the available product range. Downsides include limitations on the accuracy and available precision (which are a function of the fixed filament thickness), and shrinkage in the final work product that is both common and unpredictable. It is important to recognize FFF is subject to anisotropy (the varying properties of an item that are a function of its orientation and direction): the strength of a finished part is a function of the strength of the bonds between layers rather than of the build material itself.4 Thermal (and often humidity) control of an FFF build is also important to prevent warping from differential cooling rates across different axes of the build.
Often an FFF part will have an interior structure that intentionally has voids. This infill reduces both consumption of filament and build time. Ten percent infill is used if a part is being prototyped for fit, for example, 20 percent is a common build tradeoff, and 80 percent infill is used to maximize strength. Many infill patterns are available, and control over their density and other parameters is often left to the designer.5
Both industrial and most home 3D printers fall into this category. Other variants of material extrusion deposit liquids or slurries rather than melted thermoplastic. (An example is robocasting, which uses a ceramic slurry that does not usually require support structures because the material sets up rapidly.6) Some of the often experimental machines that layer ceramics, composites, metal-filled clays, electronic circuits, concrete, food, and living cells in a hydrogel suspension use the same basic technology but will not be discussed here.7
Two major advancements on extrusion-based printing, both at MIT spinouts, bear mention. First, MarkForged initially focused on impregnating polymer filament with chopped carbon fiber, which reduced weight, increased stability, and improved strength of the printed items. More recently, the firm separated the sintering oven from the build chamber and uses polymer filaments infused with metal to feed the extruder. (Sintering is a post-processing step in which heat bakes off impurities and in this case shrinks the built part to size; it refers to the heating [sometimes with a laser] of particles to a jelly-like state where they fuse without melting.) Workplace safety is dramatically improved: material handling in powder-bed systems can be dangerous and requires extensive precautions.8 Second, Desktop Metal is attempting to bring metal 3D printing to office spaces and other environments that lack safety hoods, advanced fire suppression, and other industrial features. No lasers or powders are involved; rather, metal rods bound with a wax and polymer interface material are extruded to build up the near-net shape. The build is then moved to a second machine called a debinder, where the proprietary interface material is removed. Then the build moves to the sintering furnace where further binder is removed and the metal consolidates to its final density, in the 96 to 99.8 percent range.9
Both companies’ metal printers are just coming to market in 2018 but bear watching as yet another blurring of the ISO/ASTM categorization.
Beginning in the early 1990s, 3D printing expanded from thermoplastic and photosensitive resins to metal fabrication. Early interest was shown by aerospace and defense contractors. Until recently, the filament model did not apply to metals. Instead, all powder-bed techniques begin with a bin of powder (either thermoplastic or metal) that is heated to near the material’s melting point then subjected to a beam of powerful energy. The particles fuse, the build platform is lowered precisely, and a new layer (fractions of a millimeter thick) of powder is spread over the top of the bin, usually with a wiper. The process is then repeated, with each layer representing a slice of the CAD drawing that has been translated into software driving the print head. The unsintered powder supports the build as it takes shape and often can be recycled. Both during and after the build, a carefully managed cooling process is critical to help prevent warping and other thermal distortions. Furthermore, careful handling of the powder feedstock is essential for health and safety reasons.
| Extrusion (Fused Filament Fabrication) | |
|---|---|
|
Method of binding |
Thermal heat |
|
Selected materials |
Thermoplastic filament |
|
Layer thickness |
.127–.5 mm |
|
Support structures required? |
Yes |
|
Selected vendors |
MakerBot, Stratasys |
|
Advantages |
Low cost, widely available |
|
Limitations |
Build chamber size limit, low resolution |
SLS technology was developed in the late 1980s and the first commercial machine was shipped in 1992. Both EOS (formally Electro Optical Systems) of Germany and 3D Systems in the US manufacture these machines. It uses a bed of powder both as feedstock and support for the build. (SLS refers to the fusing of plastics, glass, and ceramics. DMLS refers to metalworking machines, which constitute the largest installed base of any metal-capable additive technology. SLM is similar but works only with pure metals with a single melting point. The latter two will be grouped together for simplicity.) Both polymer- and metal-capable technologies build structures that are self-supporting, opening the way to complex internal geometries. Hollow ducting can be precisely engineered and rendered, for example. The advantages of powder-bed fusion systems include precision and strength of finished parts (as with dental and orthopedic appliances), the flexibility of many available materials, the lack of support pieces for overhangs and the like, and no need for post-build curing. Powdered metal is considerably more expensive than solid stock: one stainless steel variant costs up to $450 per kilogram, five to ten times the cost of bar stock. Other downsides include high power consumption, the need for measures to relieve thermal stress that might include supports, relatively rough surface finish, and a small build envelope (22′′ × 22′′ × 30′′ at the largest) relative to the large enclosure of the machines, in part because the build chamber must be filled with inert gas to prevent explosion of certain powder particles. The recommended space required for the machine cited above, costing more than $250,000, is 15′ × 15′ × 9′.
| Laser Melting (Direct Metal Laser Sintering, Selective Laser Melting) | |
|---|---|
|
Method of binding |
Laser heat |
|
Powder bed? |
Yes |
|
Selected metals |
Steels, Inconel, cobalt, some aluminum alloys |
|
Layer thickness |
.02–.08 mm |
|
Support structures required? |
Yes (thermal) and self-supporting |
|
Selected vendors |
EOS, Concept Laser (GE), 3D Systems, Renishaw |
|
Advantages |
Strength, precise detail |
|
Limitations |
Low speed, power consumption |
EBM technology was developed later in the 1990s, was commercialized in 2001, and uses electrons rather than the photons to heat the powder. A leading vendor of these systems is Arcam, out of Sweden, now part of GE. A large build envelope is in the 10′′ × 10′′ × 15′′ range; it is housed in a machine roughly 6.5′ × 3.5′ × 8′ weighing about 3,500 pounds. Successive layers of preheated metal powder, each roughly 0.0004 inches (10 microns) thick, are melted in a vacuum under the direction of a computer interpreting the build file. After each layer is complete, the build platform is lowered, a thin layer of new powder applied, and theprocess repeated. As with DMLS and SLM, parts are built up from a solid metal base plate and must be separated after the build cools, typically with a cutting tool. Advantages of EBM include the absence of voids and high strength of the finished product, high accuracy (including limited shrinkage), medium levels of surface finish, and relatively rapid build speed. The downsides include the need to work only with conductive materials, relatively large minimum feature size (typically 0.1 mm), the complexity and criticality of maintaining the vacuum chamber—in part to limit gamma ray radiation—and high power consumption.
Sheet lamination is exactly what the name suggests: successive layers of paper, plastic, foil, or other flat materials are cut with great precision either with a blade or a laser, then fused together. For the most part, the process produces aesthetically pleasing builds that can be used as architectural models, for example, but lack strength for use as functional parts. The company that pioneered the technology (Helisys) has exited the market, but such vendors as Cubic and Mcor are still active. Fabrisonic uses sheet lamination to join metal foils via ultrasonic welding; some systems include CNC machine tools built into the overall unit, and the company claims build envelopes up to six feet square. Sheet lamination is used mainly in niche applications such as architectural models built from paper.
| Electron Beam Melting | |
|---|---|
|
Method of binding |
Electron beam heat |
|
Powder bed? |
Yes |
|
Selected metals |
Copper, steels, titanium, nickel |
|
Layer thickness |
.05–.2 mm |
|
Support structures required? |
Yes |
|
Selected vendors |
Arcam (GE) |
|
Advantages |
High strength, speed relative to DMLS, high resolution |
|
Limitations |
Build chamber size limit, power consumption, vacuum chamber |
Binder jetting technology dates to the early 1990s and was patented as “3D printing.” Leading vendors include ExOne for metals, and 3D Systems and Voxeljet for sand, gypsum, and other materials. The process joins powder with adhesive binders rather than heat; the potential combinations of binders and materials (which can range from chalk to metal) are quite extensive. Further, the lack of a need for heat in the fabrication phase can open larger build volumes, and warping and shrinkage are minimal, even in large metal parts, except when post-production infiltration or sintering are performed. Binder jetting has also been used for sand molds (to receive molten steel) in the 70′′ × 40′′ × 16′′ range. Direct printing of sand molds with complex features is a key use of the technology, at multiple size ranges.
After layer-by-layer fabrication, which can include secondary material such as color from an additional print head, the part can be sintered in a furnace, which also burns off the binder. Molten bronze or other infiltrants can fill voids via capillary action, if necessary, to obtain full material density; isostatic pressure can also be employed to increase densities in metal builds. The technology is often used in aesthetic applications such as architectural models, packaging, and ergonomic verification. Functional metal parts can also be produced at relatively low cost. Advantages of binder jetting include much higher build speed than is possible in laser systems, the large build envelope, which has been used to produce room-sized architectural structures, cost that can be ten times lower than powder-bed methods, and extreme design flexibility (support structures are typically not required). The technology’s downsides include the large size of the fabrication machinery, and the limited strength and other material properties of finished parts relative to powder-bed methods.
| Binder Jetting | |
|---|---|
|
Method of binding |
Binding agent (adhesive) |
|
Powder bed? |
Yes |
|
Selected metals |
Bronze, sand, stainless steel |
|
Layer thickness |
.09 mm |
|
Support structures required? |
No |
|
Selected vendors |
ExOne, Voxeljet, 3D Systems |
|
Advantages |
Wide selection of materials, relatively lower cost |
|
Limitations |
Highly porous surface, metal parts are not as strong as laser or EBM |
Material jetting, commercialized in Solidscape, Stratasys PolyJet, and 3D Systems MultiJet machines, uses hundreds of tiny nozzles to spray layers of liquid photopolymer (in the case of Solidscape, wax for investment castings, which is not cured) onto a build tray, where they are cured and solidified with ultraviolet light. Support structures are fabricated of a dissolvable material that is removed during post-processing. A typical build size is 15′′ × 10′′ × 8′′. A wide range of materials can be used, and multi-material fabrication is possible, but resin cartridges are usually proprietary and expensive, costing up to $1,000 per kilogram. Because builds are solid rather than infilled with lattice, material usage is higher than with FDM or SLA, raising costs further. Medical models are one application of this capability: nerves, veins, arteries, and so on can be rendered in different colors, or built of transparent (or selectively transparent) materials.
More recently, electrically conductive nanoparticles have been embedded in the polymers, and a company called Nano Dimension is commercializing the ability to 3D print electronic circuit boards for rapid prototyping.10 Optomec’s aerosol jetting allows circuit elements (such as resistors, conductors, and semiconductors) to be printed on a variety of substrates.11 Elsewhere, the Israeli firm XJet has taken material jetting to the nano scale (using heat rather than light for binding metal-impregnated liquid) for both ceramics and metals, opening new markets for additively produced structures.12 As with MarkForged and Desktop Metal, XJet’s treatment of metal feedstock avoids many of the safety considerations required in powder-bed systems.
Advantages of material jetting include extreme design flexibility given multi-material and full color capability, smooth surfaces comparable to those produced by injection molding (the layer height can be as small as 16 microns), and thus very accurate visual and haptic prototypes. Production parts are not commonly produced in this fashion. Because the build does not undergo sharp thermal transitions, warping and shrinkage immediately after the build are minimal. The downsides include the high cost of the technology, the fact that material jetted parts are photosensitive and thus degrade over time, and poor mechanical strength. Post-processing is usually focused on dissolving the support structures, minimal sanding, and the application of various coatings.
In the late 1990s, researchers at Sandia National Laboratories were able to fuse metal with a laser outside a powder bed, and the technique was commercialized in the United States by Optomec beginning in 1997. The technique is known under a number of names including laser engineered net shaping (Optomec’s LENS), direct metal deposition, and 3D laser cladding. Metal powder or wire is fed onto a surface (either a piece being repaired or an irregularly shaped new build) and melted with a laser or electron beam. The technology has the advantages of larger build envelopes than powder beds can accommodate and the cost savings of creating or repairing expensive, complex metal structures: the layers can be deposited freeform on curved or irregular surfaces rather than only on a planar build platform. Downsides include a limited installed base, limited number of available materials (including some plastics and ceramics), and a rough surface texture that typically requires post-processing.
| Material Jetting | |
|---|---|
|
Method of binding |
Heat, binder |
|
Selected materials |
Photoresins |
|
Layer thickness |
0.08 mm |
|
Support structures required? |
Yes |
|
Selected vendors |
Solidscape, Stratysys, 3D Systems |
|
Advantages |
Excellent surface finish, full color, multi-material builds |
|
Limitations |
Brittleness in finished builds |
| Direct Energy Deposition | |
|---|---|
|
Method of binding |
Laser or electron beam |
|
Powder bed? |
No |
|
Selected metals |
Cobalt chrome, titanium |
|
Layer thickness |
0.089–0.203 mm |
|
Support structures required? |
No |
|
Selected vendors |
Optomec, Trumpf |
|
Advantages |
Can be used to repair parts, able to build irregular shapes |
|
Limitations |
Slow speed, requires post-processing such as polishing |
A major change came to the 3D printing market after 2014, when HP introduced first multi-material and then (in 2018) metal-based additive technology. Claimed to rely on 5,000 HP patents, the company’s Multi Jet polymer technology combines elements of other techniques: binder jetting, material jetting, and powder-bed fusion. Like SLS, Multi Jet is a powder-based technique, but no lasers are involved. Instead (similar to binder jetting), a fusing agent is jetted in combination with thermal heat to melt particles, and a detailing agent is used at object contours to improve part resolution through a similar melting process aimed at surface quality rather than strength. The resulting parts are claimed to be built up to ten times faster and 65 percent less expensive, per part, than FDM or SLS done in industrial-grade machines.13 Furthermore, HP has taken a systems perspective on additive manufacturing, addressing the process from end to end. HP’s solution begins with computer data formats, includes in-process quality monitoring and control, establishes an open-source community model for new material identification and commercialization, and integrates a cooling chamber on the printer chassis to increase speed still further. In addition, HP has instantiated an integrated model for packaging, workflow, and reuse practices related to powder handling.
HP is a large firm in a small market: its market capitalization of roughly $36 billion in 2018 dwarfed those of Stratasys (barely $1 billion), EOS (privately held; revenues in the 500 million euro range), or 3D Systems (roughly $1 billion). After its investments in Arcam and Concept Laser, GE has become the other major player in the market. HP’s scale and aggressively comprehensive approach will bear watching, as will its ability to partner with similarly large global players. A major announcement relative to an expansion into metal-based printing in 2018 further reset the market. Now that additive manufacturing has the backing of two multibillion-dollar firms, we should expect major new product introductions and possibly further mergers of existing players.
Heading into 2020, 3D printing is able to precisely deposit more materials, at more scales, and faster than ever before. Long-standing problems around build volume, material-handling safety, build planning, and limited commercial markets appear to be seeing solutions. From nano-structures to buildings, and from circuity through biology, the horizons of commercial capability are expanding. Despite the relative youth of the field, additive manufacturing has achieved many successes in both consumer and industrial markets, and it is to those markets that we now turn our attention.