In this article, it will be attempted to present a quick overview of the different generative processes. Since this area is growing and many processes are still not fully matured, only brief descriptions can be presented. However, the process stereo lithography was the first GMP to be developed and commercialized by 3D Systems Inc. and will be discussed in more details.

The presentation of the various processes will be attempted according to the scheme of shape generation. Thus, first the processes following two-dimensional layer-by-layer technique will be discussed. This will be followed by a brief discussion of the direct three-dimensional shape generation processes.

Generative Manufacturing Process and its Types

Type # 1. Stereolithography (STL) with Photo Polymerization:

This process is based on curing of a liquid photopolymer (or monomer) by an ultraviolet laser beam. The generating vat contains a UV sensitive liquid photopolymer. An NC drive controls the height of the platform on which the work is generated. Initially, it is positioned along the top surface of the liquid. A UV laser beam is reflected on to the surface of the liquid with the help of a mirror mounted on a pair of orthogonally scanning galvanometers whose positions are controlled by a computer.


The computer controls the movement of the mirror so that the beam traces the desired path on the liquid surface while generating a particular cross-section of the part being produced. The beam cures the liquid to produce the solid layer with a depth of a few tenths of a mm which is equal to the thickness of the particular cross-section (i.e., the thickness of a layer between two consecutive slices). Once the first layer is cured, the platform is lowered by a distance equal to the thickness of a layer.

Then, the laser beam scans the liquid surface again to produce the next layer. The process is repeated till the topmost cross-section of the object is generated. Next, the part is removed from the vat and the excess material is removed from the crevices and openings using ultrasonic cleaning technique. An alcohol bath is used to clean any unused polymer. To solidify the trapped liquid polymer inside the hatched pattern, the part is subjected to post curing operation that is carried out by applying intense long wave UV radiation.

In this process, solidification of the liquid resin is accomplished through a process called photo polymerization which links small molecules, called monomers, into large molecules, called polymers. This also results in major changes in the bulk properties. Since the loose van der Waals interaction among neighbouring monomers is replaced by a network of covalent bonds, the shear strength increases significantly, changing the liquid to solid. The polymerization process is very energy efficient as the process is exothermic.

This allows low power UV lasers in stereo lithography process. In comparison to powder sintering type of operation, stereo lithography requires about 1000 times less power! Upon exposure to UV radiation, free radicals (R•) are produced from the photo-initiator (I) when the photons of certain frequencies are absorbed. The free radicals react with a monomer (M) and the chain reaction that starts is –

Solidification can take place both in a point-by-point fashion and curing lines at a time. In the case of low power laser beams, the beam scans the resin surface so that a series of voxels (volume picture cells) get solidified as shown in Fig. 7.51. The size of the voxels should be adequate to ensure connection with the neighbouring voxels and each layer should also be connected with the layer solidified prior to the current one as indicated in Fig. 7.51.

When the power of the laser beam is low, voxel formation is achieved by a point-to-point NC control of the mirror that causes the laser beam to stop at each voxel point. The beam need not be switched off in between voxels and the traversing speed being high, polymerization does not take place during this period. When high power lasers are used, continuous lines can be cured, forming a solid parabolic cylinder as indicated in Fig. 7.52.

The degree of solidification through photo polymerization depends on the number of photons impinging on to a particular target volume, i.e., the dose of radiation received. Solidification requires a minimum critical dose. The surface energy received while traversing a line can be expressed as –

Where P is the power in watts and v is in cm / sec.

The strength of a photo polymerized resin depends on the total exposure. Young’s modulus Y, which can be an indicator of the mechanical strength of the solidified resin, is a function of the energy E. However, it is obvious that when E < Ec, there will be no solidification and Y= 0. When E exceeds Ec, Y(E) increases sharply and approaches a limiting value Ymax asymptotically. When

E ≫ Ec, Y= Ymax. On the other hand, if E is not much larger than Ec, then


Y = k(E – Ec),

where k is a material constant with the dimension of length-1 and its numerical value is in the range (1.5-9) x 105 m-1 when E is in mJ / cm2 and Y is in N / m2. Table 7.10 gives the important properties of a few resins used in stereo lithography.

It is possible to develop a process of shape generation, somewhat similar to the STL process we have described, in which a resin is not solidified through photo polymerization but a thermosetting liquid polymer is selectively solidified by heat. Quadrax has marketed a system of this type in which a 5-W Ar-Ion laser is used. However, this company ultimately did not succeed in capturing the market for a number of reasons, patent infringement and litigation being one of the main reasons.

Type # 2. Selective Laser Sintering (SLS):


This process was first conceptualized and developed in 1987 at the University of Texas at Austin; the process was commercialized and marketed by DTM Corporation, USA. In this process, a part is created following a layer-by-layer approach by fusing powdered thermoplastic materials with a high power laser beam. This process of getting powder grains bonded together by localized partial melting is called sintering.

In general, sintering occurs when a particle’s viscosity drops due to higher temperature and the surface tension effect overpowers the viscosity. SLS can be of two basic types, viz., indirect SLS and direct SLS. In the case of indirect SLS, the metal or ceramic powders get bonded together through a polymeric material which softens (or melts) and forms necks between adjacent grains. The binding takes place through polymer-polymer bonds.

In the indirect SLS, either metal (or ceramic) powders are mixed with a small amount of polymer or metallic (or ceramic) powder grains coated with polymer. In the direct SLS process, laser beams of much higher energy are used to partially melt the metallic powder and get sintered.

Figure 7.53 shows the nature of indirect and direct SLS products. Because of the very nature of the solidification process, the products are generally porous and of low density. (In the case of indirect SLS, the density is lower than that achieved through direct SLS which can be of the order of 80-90%). The density produced by indirect SLS is 80-90% of that achieved through direct SLS.

The layers formed by the sintering process reside within the powder and no separate support structure is required. This, to an extent, simplifies the planning and designing task. When all the layers are formed, the product, embedded in the loose powder, is removed and cleaned.

The machine marketed by DTM Corporation consists of a cylindrical container in which the parts are generated. There are two powder supply cartridges on both sides, a counter-rotating roller for laying fresh layer of powder on the top of the newly created layer of the part, a 50-W CO2 laser and its associated optics along with a computer controlled mirror to guide the laser beam. The basic scheme of the machine is shown in Fig. 7.54.

The part building chamber is purged with an inert gas, generally nitrogen. The chamber is also maintained at an elevated temperature so that the powder in the top layer is just below the sintering temperature. This helps to minimize the power requirement of the laser beam, and also increases the speed of sintering process.

In the SLS process, a thin layer (a few tenths of a millimeter thick) of powder is spread into a cylindrical part building chamber with the help of a counter-rotating roller which moves from one side to another of the cylindrical chamber flanked by two powder feed cartridges. The powder feed pistons move up with the consumption of powder in creating each layer and keep the powder in the supply cartridges always at the required level. The top layer in the building chamber is raster scanned (closely spaced parallel lines) with a 50-W CO2 laser beam.

The requisite area, representing the part geometry at that particular layer on which the laser beam scans through, fuses together and solidifies. Since the chamber is maintained at an elevated temperature, it not only reduces the power requirement of the laser beam but also helps to keep thermal shrinkage during fabrication low keeping the part distortion minimal. After every layer is sintered, the part supporting platform is lowered by the equivalent of one layer thickness and a fresh coat of powder is laid.

The process is then repeated. The un-sintered powder remains to support the developing part as a form-fitted ‘cake’. Thus, the part is well stabilized during the building process as already mentioned. After all layers are built, the supporting piston is raised; the un-sintered powder mostly falls off the part and a spatula is used to remove the additional powder. The un-sintered powder can be reused. A layer of un-sintered powder, about 25 mm thick, is left covering the part that serves as an insulation which helps to reduce distortion as the part cools. The coating is removed by brushes and air jets after the cooling is over.

The parts produced by SLS are of low density as no compaction pressure is applied unlike that in the conventional powder metallurgy based processes. The parts produced by sintering polyvinyl chloride have a relative density of only 60% (i.e., 40% of the part volume is air). The materials used for SLS include PVC, polycarbonate, investment wax, nylon, and ceramic and metal powders. One distinct advantage of SLS is that different materials can be used while building a single part.

Almost all generative manufacturing processes produce a vertical stair-step surface finish since parts are built layer upon layer. The surfaces of parts produced by the SLS process suffer additionally from roughness problems. One major source of this roughness is the granular nature of the raw material which are powders with grain sizes in the range 80-120μm. Besides, the raster scan laser beam drawing also results in horizontal stair-step effect as explained in Fig. 7.55.

Some improvement of surface finish may be possible by rotating the orientation of raster by 90° at every alternate layer. This will distribute the roughness evenly on all surfaces. Further improvement of surface finish is possible by outlining each cross-section prior to the drawing of rasters, though this will increase the build time.

Type # 3. Selective Powder Binding (SPB) or Three-Dimensional Printing:

Selective powder binding (also called three-dimensional printing) was originally developed at MIT in the early 1990s. The process is based on creating a solid object from a refractory powder by selective binding through the application of a colloidal liquid silica binder. The liquid binder is applied selectively to thin layers of powder, in the form of droplets, using the inkjet technology causing the particles of powder to stick together wherever the binder droplets are applied.

Thus, it has certain similarities with SLS. Soligen licensed the three-dimensional printing technology patents from MIT for direct shell production casting. Z Corp also procured license of the three-dimensional printing technology for building models.

The inkjet technique has also been used to develop solid objects by directly injecting material droplets at required locations. Either one or two inkjets can be employed for depositing tiny drops of hot liquid thermoplastic materials. More than one material can be used for making the part if two inkjets are used. Usually, one jet is employed for creating the support structure and the other forms the main part.

With a different material, the support structure can be easily removed after the part formation is complete. Solidoscape commercialized such inkjet machines.

In the SPB type three-dimensional printing process, a thin layer of ceramic powder is laid on a flatbed (Fig. 7.56). Next, a fine jet of ceramic binder is ejected onto the powder at locations where solidification is desired. The application of binder droplets can be done in two different ways. In the drop on demand technique, a droplet is ejected by the inkjet mechanism when a drop is needed while the inkjet mechanism traverses the layer of powder.

In continuous jet systems, the droplets are ejected continuously while the nozzle traverses the layer surfaces. But at locations which are not to be solidified the inkjet is diverted by an electric field. The droplets are electrically charged while leaving the nozzle tip. The nozzle is moved across the powder surface in a raster scan while the computer generated electrical signals control the deposit of the binder.

After the completion of the selective binding operation of the powder layer, the platform supporting the part is lowered by one layer thickness and the cycle is repeated. Figure 7.56 shows the various stages of the process, including the completed part. The inkjet based print head consists of an array of a large number of jet nozzles, each one capable of operating at few tens of kHz.

Typically, for a layer of 0.5 m x 0.5 m, it takes about 4 sec to complete the cycle using the drop on demand technique. When continuous jets are used, the time for solidification of the layer can be much smaller and even as low as a fraction of a second.

After all the layers are completed, the part is cured at 120°C for about two hours. The unbound powder is then removed. For ceramic parts, a final firing at 1000-1500°C is required to impart the object its full strength. The typically used powders are aluminium oxide, zirconia, zircon, and silicon carbide. The minimum feature size is about a fraction of a mm. Difficulty in achieving good surface finish is one of the major problems of the process. The removal of unbound powder from narrow passages and enclosed cavities also poses difficulties.

Type # 4. Ballistic Particle Manufacturing (BPM) – Layered Process:

This technique for creating three-dimensional solid objects from the CAD model was developed by Perception Systems Inc. It involves shooting of droplets of molten material at required positions. As in the selective powder binding process, here also material is supplied through an array of drop on demand inkjet ports. Molten wax droplets of about 50μm diameter are ejected at the rate of 12,500 drops per second.

Unlike most other generative manufacturing processes, BPM is possible both for layer-on-layer fabrication and direct three-dimensional shape generation. The two-dimensional layer-on-layer process is based on generating layers from wax droplets. In this process, the CAD model is developed for both the part and the support structure. Figure 7.57a shows the CAD model of a part and Fig. 7.57b the CAD model of the part-cum- support.

The part is generated from wax whereas the support is developed from polyethylene glycol (Figs. 7.57c and 7.57d), a synthetic wax that is soluble in water. The deposition of the part and support material is accomplished by sorting droplets from an array of 32 piezoelectric inkjet ports operating at 10 kHz. On contact with the previously generated layer, the hot droplets momentarily melt the contact surface of the previous layer. On subsequent cooling and solidification, a homogeneous material is formed of the desired shape.

After the completion of deposition of all layers, the object is placed in warm water bath to dissolve the support material, leaving the desired part. The accuracy of the process depends on the accuracy of the position of droplets, which in turn is dependent on the accuracy of the location of the piezoelectric jet system and the ballistic paths of the individual droplets.

Thus, it is desirable that the jet ports be as close to the substrate as possible. The layer thickness is monitored by a feedback loop using proximity sensors for measuring the distance between the jet ports and the substrate. In the commercially available systems, parts have been generated with 90μm layer thickness.

Type # 5. Fused Deposition Modelling (FDM):

This is one of the more popular methods for generative manufacturing. In this process, three-dimensional objects are produced by depositing a molten thermoplastic material layer by layer. Stratasys Inc. is commercially manufacturing machines for FDM. A solid filament of thermoplastic material with 1.25 mm diameter is fed into an x-y controlled extrusion head. The material is melted by a resistance heater at a temperature of 180°F (1°F above its melting temperature).

As the head is moved along the required trajectory using computer control, the thermoplastic material is deposited by extruding it through a nozzle by a precision volumetric pump. As the extruded material, deposited as a fine layer, comes out with a temperature just above the melting point, it re-solidifies within 0.1 second by natural cooling.

To ensure proper adhesion of the deposited fused material to the previously deposited layer, the object temperature is maintained just below the solidification temperature. After one layer is deposited, the platform, supporting the object, is lowered by one layer thickness.

To maintain stability in the process, the rate of flow of the extruded molten filament is controlled to match – (i) the travelling speed of the depositing head (which can go upto 380 mm / sec), (ii) the desired thickness of the layer (that varies from 0.025 mm to 1.25 mm), and (iii) the width of the deposited line (which varies from 0.23 mm to 6.25 mm). The repeatability and positional accuracy of this process are claimed to be about ±0.025 mm with an overall tolerance of 0.125 mm over a cube with 305 mm sides.

The FDM process is still not very suitable for parts with very small features. The typically-used materials for the process include investment casting wax, wax filled adhesive material, and tough nylon-like material. Polymer type thermoplastics can also be used.

Type # 6. Laminated Object Manufacturing (LOM):

In this process, parts are produced by successive bonding of layers of sheet materials (mostly paper type material) and laser cutting of each cross-section. Thin sheets of plastics and composites are also used. The part acquires a wood-like structure and quality. Helisys Inc. released the first commercial model of LOM system. Sheet material is supplied from a roll and the unused portion of the material is wound up at the take-up roll.  

The sheet is coated with a heat sensitive adhesive. When the fresh material comes over the work table, a heated roller presses it down to the uppermost layer of the object being fabricated. Then, a CO2 laser beam cuts the outline of respective cross-sections. The beam intensity and speed are so adjusted that only one sheet is cut. The sheet material outside the desired cross-section is cross-hatched by the beam into squares (called tiles) so as to separate the part easily after it is generated.

After the cutting of one layer is completed, the platform is lowered by one sheet thickness and the rollers supply a fresh area of the sheet material. Interlayer adhesion near the boundaries is a problem for LOM. This problem is partially reduced by a method called burn out. The area on the previously laid layer where gluing is undesirable is cut with a tightly spaced cross-hatched pattern. A hollow part cannot be generated by LOM as the excess material remains trapped inside.

Type # 7. Solid Ground Curing (SGC):

In this process, whole layers are simultaneously cured according to the required cross-section. Cubital marketed an SGC system called Solider. A thin layer of liquid photosensitive resin is applied and then exposed to a strong UV radiation through a mask. The transparent areas of the mask correspond to the desired cross-section of a particular layer. The UV radiation solidifies the exposed areas of the resin and the uncured liquid resin is removed from the unexposed areas.

These areas are replaced by wax to build up the support structure. Finally, the cured resin and the deposited wax are both machined to a predetermined thickness using an end mill cutter. The cycle is then repeated till the complete object is formed. Concurrently, another cyclic operation is performed to prepare the mask for each layer. A glass plate is ionographically charged to create a pattern according to the required cross-section, which is then developed by using a toner, as done in photocopying units.

Once the curing of the required cross-section is over, the pattern from the glass mask is erased and it is used for creating the mask for the next layer. A computer containing the part geometry controls the charge pattern. Typically, the charge is deposited on the glass plate in raster lines with 11.8 lines per mm.

In more advanced versions, 40 lines can be accommodated per mm. The layer thickness ranges from 0.03 mm to 1.27 mm.  

High viscosity low shrinkage resins can be used for the SGC process. Since curing is done of a whole layer, the build time per layer is independent of part geometry, and it is about 50 seconds. An accuracy of 0.03 mm can be achieved in a 25 mm part dimension.

Type # 8. Direct Three-Dimensional Generative Manufacturing Processes:

Though most of the commercially available generative manufacturing process units fabricate parts layer by layer, it is possible to build objects directly in three-dimensional space. In all layer-by-layer methods, the lower layers have to be created before the next layer is deposited. But direct three-dimensional techniques do not require to create the lower portions first.

Direct three-dimensional techniques can be of three types as follows:

i. Shape generation through point by point.

ii. Shape generation through surface by surface.

iii. Shape generation through simultaneous creation of the whole object. Some techniques discussed now are still at the developing stage.

A. Beam Interference Solidification (BIS):

This method was patented by Formiographic Engine. The material which is used in the process is a photosen­sitive transparent liquid plastic (monomer). When the monomer is subjected to a laser beam of a particular frequency, it reaches a reversible metastable state and no bonding reaction takes place.

But when a part of the liquid that is already in a metastable state is hit by another laser beam of a specific (but different) frequency, polymerization of the metastable state takes place, resulting in the solidification of a voxel (volume picture cell) at the intersection of the two beams.  

By moving the two laser guns in a particular way, the volume of the desired object can be generated voxel by voxel. In spite of the elegance of the concept, there are a number of serious difficulties for its practical application. The intensity of the beam decreases continuously during its passage through the resin because of absorption.

This poses a serious difficulty in programming the laser beams to maintain uniformity in the characteristics of all voxels. The problem is further compounded because of shadow effects produced by the portions of the object already solidified.

B. Ballistic Particle Manufacturing (BPM):

The process has already been dis­cussed in its layer-by-layer application. In the direct three-dimensional approach, the part building is achieved by shooting of molten droplets on the top of each other. Two piezoelectric inkjet printing nozzles are guided by manipulators to deposit the molten droplets according to the need from any direction instead of using an array of such inkjet nozzles as done in the layer-by-layer approach. Six-axis robots can be used for the purpose. One important advantage of this approach is the elimination of support structures.  

C. Direct Metal Deposition (DMD):

The University of Michigan, Ann Arbor, has commercialized a laser based machine for generating three-dimensional metallic parts. Metal powder is supplied which is melted by a high power laser beam and functional metallic part can be generated. The deposition head is guided by a manipulator.

D. Holographic Interference Solidification (HIS):

This exotic process is also based on photo polymerization of photosensitive resins. But in this process, the part is not created voxel by voxel; instead, a three-dimensional image is projected in a vat containing a photosensitive liquid monomer and a whole three-dimensional surface gets solidified as a whole. The holographic film for projecting the image is created with a CAD system. A system based on this principle has been developed by Quadtec Pty., Melbourne.

The long-standing desire of manufacturing engineers to produce solid parts directly from the design data stored in a computer seems to be on the verge of realization. Though the current methods of generative manufacturing are used for rapid prototyping, primarily, the trends of development indicate that soon these processes will be used for production of functional parts. Some indirect benefits will also be there once functional parts are produced by these techniques.

At present, products have a limited life mainly because of the cost of spare parts and their non-availability after some time. Keeping a large stock of spares for an indefinite period of time means blocking of capital. But when generative processes will mature enough to produce spare parts, spares can be stored electronically eliminating the blockage of capital.

The major advantages of the generative manufacturing processes can be summarized as follows:

i. No tools and fixtures are required.

ii. No restriction on geometry of the part shape exists.

iii. Composite parts and assemblies can be produced in one go.

iv. The process is very suitable for computer integration.

v. Small batch production of complex parts is economically viable.

The generative processes will also play a very major role in manufacturing micro parts. These processes are also responsible for bringing bottom-up approach in the manufacturing scenario and can be considered to be the forerunner of the ultimate process—manufacturing by self-assembly of material.