STEREOLITHOGRAPHY, SELECTIVE LASER SINTERING AND POLYJET: EVALUATING AND APPLYING THE RIGHT TECHNOLOGY

Rapid Prototyping (RP) is a powerful tool for reducing time-to-market while improving quality and reducing cost. When one has decided to take advantage of the benefits of RP, the challenge is selecting the right process for the task. Choosing between the various RP technologies can be difficult, and it has become even more challenging as the technologies have matured. Strengths and weaknesses are inherent in all processes, and RP is no exception. To properly select the RP method for a specific application, one must understand the limitations as well as the strengths of each technology. Without hands-on experience, many of these factors will be obscure. Even with a full understanding of these parameters, one will find that there are trade-offs between these technologies that cloud the decision-making process. Three leading rapid prototyping technologies in use today are Stereolithography (SLA), Selective Laser Sintering (SLS) and PolyJet. As a high volume user of all three technologies, the author of this article is well-qualified to offer an unbiased description of the capabilities of each technology, and this article will describe and compare the processes as well as provide the necessary guidelines that should be considered in choosing the right one.

Figure 1: Illustration of the test part used to compare the accuracy of SLA, SLS and PolyJet

Choosing the Right Process

There are numerous issues to consider when selecting a process for a given application. These factors are grouped into three categories: the physical considerations, operational considerations, and application considerations.

The assumed goal is to receive a prototype quickly, accurately and cost effectively. To achieve this aim, one should evaluate these factors with equal weight. It is also important to consider all aspects of the technology combination with the intended material to be used for the prototype. Selecting a technology without considering the material will yield questionable results.

Process Overview

Stereolithography (SLA): Developed by 3D Systems, SLA is the most widely used RP technology. SLA enables the creation of complex, 3D models by successively “laser-curing” cross-sections of a liquid resin. A UV laser contacts the resin, which is a photopolymer, causing the material to solidify. Although SLA is limited in its range of applicable materials, it is widely used for conceptual visualisation, form and fit analysis, pattern creation and light functional testing.

Selective Laser Sintering (SLS): Developed by DTM (now owned by 3D Systems) SLS is widely used for funcional applications. The SLS process creates 3D objects, layer by layer, from powdered materials. Heat from a CO2 laser fuses (sinters) the powder within a precisely controlled process chamber. A major distinction between SLS and other RP technologies is the wide variety of materials that can be utilised. Many of these materials prove suitable for function analysis.

PolyJet: Developed by Objet Geometries, the PolyJet technology was commercially released in 2001, when it was best known by the names of the specific equipment incorporating the technology — Quadra and QuadraTempo. PolyJet creates rapid prototypes with an ink-jet process. Instead of ink, however, the print head deposits photopolymer that is immediately cured with a UV lamp. The two most notable characteristics of PolyJet are the layer thickness (20 microns) and resolution (features down to 0.0015”). Equally impressive is the fact that PolyJet delivers this resolution with build times comparable to SLA and SLS parts built at 0.0006”. The resulting prototype requires little, if any, post-processing work for concept and form/fit models.

Physical Considerations

When evaluating RP processes, the first issue is whether the technology can provide a prototype that satisfies the physical requirements. Physical considerations relate to the quality of the prototype and how well it matches performance demands.

Material Properties

There have been significant developments in the materials available for both SLA and SLS. The advancements have improved both physical properties and operating parameters and have extended the list of appropriate applications. Developments have also increased the number of available materials. For example, SLA now has more than 15 varieties of resins from which to choose. As a new technology, PolyJet is limited in its scope of materials. Overall, SLS offers the widest latitude in material properties, and many are suitable for functional applications, SLA and Polyjet are limited to photopolymers.

SLA

Somos 11100 (WaterShed): This material has a low moisture absorption rate (0.34%), which preserves the mechanical properties when exposed to humid conditions. WaterShed is a strong material that is suitable for some functional applications.

Somos 10100 (WaterClear): This material is best known for applications that require clear parts. However, it also offers outstanding material properties when compared to other SLA resins. WaterClear parts are often used for visualising fluid flow and the internal structure of complex assemblies.

Somos 9100: This material delivers mechanical properties that approach many of those of polypropylene. These properties make 9100 a good general-purpose material. Its applications include form and fit analysis, patterns and light functional testing.

Somos 8100: This material is suited for applications that require durability and flexibility. A low flexural modulus and high percentage of elongation at break combine to deliver prototypes that are semi-rigid and flexible. Snap fits and living hinges are appropriate applications for 8100.

PolyJet

FullCure M-510T-Y: This general-purpose resin is an acrylate-based polymer and is best suited to concept modelling, form and fit review and tooling patterns. Its mechanical properties do not approach those of production grade thermoplastics.

SLS

DuraForm PA: This material is a polyamide (nylon), which makes it an excellent choice for functional applications. Approaching the mechanical properties of injection moulded thermoplastics, DuraForm PA offers exceptional tensile, flexural and impact strength.

DuraForm GF: A glass-filled variation of DuraForm PA, this material offers tensile strength and impact resistance that is superior to other materials. For functional applications that require strength and rigidity, this is a suitable material. Testing in harsh environments consisting of high temperatures or chemical agents may also demand the material properties of DuraForm GF.

Somos 201: This material produces prototypes with rubber-like properties. An elastomeric polymer, Somos 201 provides elongation of 111% and a Shore A hardness of 81 for prototyping components that are flexible. This material will also accommodate exposure to high heat and chemical agents.

LaserForm: This material is a binder coated 420 stainless steel. During the SLS process, the binder is sintered. This green part is then baked to burn off the binder and infiltrate with bronze. The resulting part has approximately 60% stainless steel and 40% bronze. LaserForm applications include the production of metal parts and rapid tooling.

CastForm: CastForm is a proprietary polystyrene-based material that provides lower operating temperatures and enhanced surface finish. These benefits, combined with low burn-out temperatures and low ash content, make CastForm an ideal material for patterns in the investment casting process.

Often, material properties are a critical part of the decision-making process. When this is true, selection can be difficult due to trade-offs in factors such as accuracy and surface finish. If material properties are the primary concern, it is best to select the appropriate material and then evaluate the capabilities of the technologies that utilise that material. It is important to take some caution when reviewing vendor-supplied material properties. Although there are testing standards, variances in build parameters, machine type and elapsed time can yield significant deviations in the results. For example, DSM Somos reports that a variation in hatch overcure can alter the tensile modulus of SLA materials by 30%. With this latitude, vendors may tend to select the testing conditions that present their material’s most favourable results.

Accuracy

The accuracy of a rapid prototype depends upon many factors, the most obvious of which are operator capability and system configuration. Other considerations include the time frame in which measurements are taken, environmental exposure and the amount of post-process finish work. Tolerance deviation also depends upon the axis along which measurements are taken. (See Figure 1 and Table 1).

Part Construction: Upon completion of the prototype, before post-processing, SLA and PolyJet provide greater accuracy than SLS. The shrinkage of the SLA epoxy resins is significantly less than that of the SLS plastic materials. The SLA epoxy materials experience less than 0.1% shrinkage during the build process, while the SLS materials yield shrinkages of 3.0–4.0%. The lower shrinkage of the SLA resins is simple to predict and easy to control. Although PolyJet has a shrink rate that is nearly double that of SLA (0.18%), the accuracy is similar. Thin build layers, fine resolution and tight control of the jetting process negate the larger material shrinkage. The larger shrink rates of SLS increase the tendency for the prototype to warp, bow or curl. Also, the SLS process is less predictable and controllable since it relies on raising the temperature of the powders to just below their melting points. Reliance on heat and heat transfer makes SLS dependent on chamber temperature, laser output and heat retention within the previously sintered powder. Should elevated temperatures be present in the unsintered powder, sintering of the part may cause undesired materials to fuse to the surface.

“Z” Axis: Consider the “Z” axis to be along the height of the part in its build orientation. In all processes, the Z-axis can have greater tolerance deviation than in the X - Y plane. Some of this inaccuracy results from the layer-additive process common to all RP technologies. During the slicing of the STL file into the desired layers, there will be round-off error. Should the top or bottom surface of a feature not be coincident with a layer, the surface height will be adjusted so that it is represented by the thickness of one layer. In this respect, PolyJet has the advantage because of its inherently thinner layers.

“Z growth” also affects dimensions in this axis. In SLA, as successive layers are cured, the beam imparts additional energy to layers below. This undesired energy can cause the lowest layers to thicken. In SLS this additive effect will also happen. However, in SLS this growth is less predictable since it is sensitive to geometry, build time and part placement in the build envelope.

Stability: Once removed from the powder cake and fully cooled, SLS parts are dimensionally stable. Unlike SLS, SLA and PolyJet parts are susceptible to additional shrinkage and creep after part construction. Therefore, a part measurement taken one week after production may vary from those taken immediately after the part was completed. Heat, moisture and chemical agents can also affect the photopolymers of SLA and PolyJet parts. Although these processes produce a more accurate part, exposure to any one of these elements can impact tolerance deviations. Environmental factors are discussed in more detail in a following section.

Post-Processing: Conventional pattern or model-making skills ultimately control accuracy. Once a file or sandpaper is applied to the prototype, the technician — not the technology — controls the tolerance. Finishing operations (post-processing) are required on both SLA and SLS parts. However, SLA is more likely to be affected because of the necessity to remove support structures from the bottom surface of each part. In this regard, PolyJet has an advantage when parts are not being prepared for a paint-ready or mould-ready finish. This is because PolyJet requires only a light pressure washing to make it a suitable, visually appealing prototype. Therefore, PolyJet part accuracy is not affected or controlled by manual labour for many applications.

Surface Finish

PolyJet provides smoother surface finishes than SLA, and SLA is smoother than SLS. However, the Ra value depends upon the surface to be measured and the amount of finishing work that has been performed. Users of SLA have generally found that the smoothest surface finish is on the top face of the prototype. The uppermost faces are contrasted by the rougher side walls and bottom surfaces. The side walls of SLA parts illustrate the striations between build layers. The bottom surface of an SLA part is affected by the support structures. When removed, and prior to a finishing operation, the supports will leave rough areas and pits on the bottom face. The preciseness of SLA that originates from the controlled and predictable nature of the process, combined with good surface finish, will allow obvious detection of stair-stepping if it is not eliminated with finishing.

Due to the sintering operation of SLS, all surfaces demonstrate rough and porous qualities. Since sintering fuses the powdered material without melting, voids are created between particles. In addition, excess heat generated during the sintering process can cause further surface finish degradation. This heat results in undesired material sticking to the part surface. Although surface roughness negatively impacts a prototype, it can return a benefit by obscuring the stair-stepping effect and side wall layer striation common to all RP technologies. The fine resolution and the thin build layers of PolyJet deliver exceptional surface finish characteristics upon completion of the build. Depending on the prototype’s application, the resultant surface finish can be acceptable without any finishing work. One exception is the bottom surface — especially for large, flat areas. The interface between support material and the bottom of the prototype often yields small pits on the part’s surface. It is also important to note that that the appearance and feel of a supported surface will be different than that of unsupported areas. The supported area will have a matte finish with a slight texture, while all other surfaces will have a smooth, glossy appearance.

SLA Ra(μin) SLS Ra(μin) PolyJet Ra(μin) Top Surface         Post Build Light Sanding   13 73 559 89 7 43 Side Wall         Post Build Light Sanding   340 90 533 203 179 87 Bottom Surface         Post Build Light Sanding   170 107 580 125 215 56

Table 2: Surface finishes associated with each technology

As illustrated in the surface finish comparison in Table 2, finishing work can be an equaliser. (See Figures 2–5). Poor surface characteristics can be greatly improved with light sanding, while glass-like surfaces will be slightly degraded. With enough finishing effort, each of these technologies is capable of producing a mould-ready or paint-ready surface. The key difference is how much time it will take to achieve the desired result.

Figure 2: SLA Top Surface (Unfinished) Figure 3: SLA Side Wall (Lightly Sanded)) Figure 4: SLS(Unfinished) Figure 5: SLS (Sealed and lightly sanded)

Feature Definition

SLA offers better feature definition than SLS. In general, SLA is capable of producing a 0.010” feature while SLS can replicate features of 0.025”. The precision of both processes impacts their ability to replicate smaller features. In addition, the spot size of the laser used in the process defines the lower limit. For most SLA systems, the laser has a spot size of 0.010”. SLS has a laser spot size of 0.018”. 3D Systems’ Viper si2 is an exception to this generalisation. Released specifically for fine detail work, the Viper has a spot size of 0.003” and can capture finer details than the rest of 3D Systems’ technologies.

PolyJet has altered the expectations of reasonable feature definition delivered with reasonable build times. The print head deposits very fine droplets of material that allow it to replicate features down to 0.0015”. Unlike the Viper, PolyJet can build prototypes with this level of detail without an increase in build time. Although the system is capable of replicating micro features, removing the support material can be problematic. In some instances, small features will not withstand the removal of the support material. For those that do survive, there is the possibility that a thin wall or tiny stand-off may not withstand routine handling.

Environmental Resistance

SLS prototypes provide material properties similar to those of the thermoplastics on which they are based. These properties include resistance to environmental exposures. The photopolymers of SLA and PolyJet offer significantly less resistance to these same environmental concerns. For prototyping requirements, the most frequent concerns are resistance to temperature, moisture and chemicals. SLS DuraForm can withstand moisture, heat up to 325ÞF and many chemicals. DuraForm parts, with a sealant applied, have been used in water-tight and under-water applications without any indication of swelling due to absorbtion. Many SLS materials are reported to accommodate exposure to chemical agents such as acids, bases, alcohol, hydrocarbons, ethers and ketones. PolyJet and most SLA parts constructed in epoxy resins should not be subjected to moisture, heat over 115ÞF or many chemicals. Although new materials have drastically reduced sensitivity to moisture, some may demonstrate swelling and changes in the material properties when exposed to high humidity or immersion. Exposure to temperatures in excess of 115ÞF will cause these parts to soften, which allows warping and distortion. Further, most chemicals will deteriorate these photopolymers. PolyJet parts react much like SLA but are less sensitive to moisture.

Operational Considerations

When one has considered the physical requirements, the next focus should be the operational parameters. All three technologies have process-related strengths and weaknesses. These considerations result from the equipment and building methods.

Part Size

The SLS SinterStation 2500plus provides a usable build envelope of 13” x 11” x 15”. The Sinterstation 2500 has a physical work envelope that is larger than these specifications, but the usable work volume is constrained by part geometry and operating parameters. The build area of an SLA 7000 is 20” x 20” x 24”, and every cubic inch of that envelope can be used, regardless of part geometry. The PolyJet Quadra and QuadraTempo have the smallest build envelopes of the three systems, measuring 10.6” x 11.8” x 7.8”. If the prototype is too large to build in one piece, it is common to split the CAD or STL file into sections so that the work envelope can accommodate each portion. Once all the sections are built, they are then bonded to form the prototype. SLS prototypes tend to be better suited for bonding. Due to the porosity of an SLS part, the adhesive penetrates the part surface to form a bond that is stronger than that found when joining SLA or PolyJet sections.

Build Times

Since the build time for each process is dependent on different parameters, general time comparisons between technologies are not always accurate. Allowing for exceptions, it is appropriate to assume that all three processes are fairly equal in typical build times. The part volume and height define build times for SLA and SLS. However, SLA tends to be more sensitive to part height. The amount of support structure can also impact SLA build times. For most parts, the time to draw support structure is insignificant. But when a large amount of support is required, SLA build times may exceed those of SLS parts. Additional consideration should be given to the material used since each has its own unique build parameters, which can significantly impact build times.

PolyJet is unique in that it is insensitive to part volume while very sensitive to the “Y” dimension. Like the other processes, build height is a key factor in determining build time. Each pass of the print head will construct a 2.5” swath of geometry. For every 2.5” of part, as measured in the “Y” axis, an additional print head pass is required. Each pass requires a fixed amount of time that is not affected by the surface area of the current layer of the part. Therefore, build times for PolyJet are simply calculated by multiplying number of passes, number of layers and time per pass. Table 3 shows build times for the accuracy test part illustrated in Figure 1. Note that the SLA and SLS parts were constructed with 0.006” layers on an SLA 7000 and Sinterstation 2500plus respectively. The PolyJet part was constructed on a QuadraTempo with 0.0008” layers.

Support Structures

SLA requires a support structure to be added to the part to successfully build a prototype. Support structures are thin ribs placed at 0.25” intervals to form a checkerboard pattern. Supports serve two purposes. First, they rigidly attach the prototype to the build platform, which anchors the part to prevent it from shifting while the platform dips into the liquid bath and the blade sweeps the surface of the resin. Second, the supports attach to any downward facing surface to fix the feature in place. Without this support, a single layer of cured material would sway or deform within the liquid environment. To complete the prototyping process, these supports must be removed. Support structures in the SLA process can impact the quality of the part. All SLA models require a hand-finishing operation to remove the supports. This post-processing can affect the accuracy of the geometry. When manually removing the support, tolerance deviations will result from the removal of too little support or too much part. In addition, material properties make the SLA models susceptible to chipping and breakage of small details. Accessibility to support structures is also a concern. Should a part have internal cavities with little or no access, the supports will remain in the prototype.

PolyJet uses a gel-like support material to create an interface between the build platform and the part; to support overhangs; and to encase and support side walls. The support material is deposited concurrently with the model material. The support material is easily removed from the part with a light pressure washing. However, due to the available material properties, small features may be difficult to preserve when exposed to the pressure washing spray.

SLS does not require any form of support structure. The powder that surrounds the sintered material acts as a fixture by encasing the prototype in a “cake”. When a build is complete, brushing, vibrating or air blasting the cake will expose the prototype to complete all required processing.

Application Considerations

What is the purpose of the prototype? One should define the intended applications and weigh the physical and operational considerations. The inherent strengths and weaknesses of each process, as described previously, will aid in identifying the right technology for the intended application.

Concept Model

Rapid prototypes are frequently used to illustrate a design to management, potential customers and others on the design team. As a visual aid, rapid prototypes excel in turning the intangible nature of engineering prints and CAD databases into physical examples of a design. In general, SLA and PolyJet models are preferred for presentation and conceptualisation. the combined benefits of surface finish, accuracy and replication of fine detail make these technologies ideal for display purposes. To mimic a production piece, these parts are often finished and painted to create a realistic representation of a product that has not yet been manufactured. With lower acquisition and operating costs, PolyJet parts are typically more cost-effective that SLA parts and therefore more likely to be used. Even in a show-and-tell environment, material properties can be a concern. When illustrating a product design, it can be unnerving when a snap fit fails. It can also be unsettling when a prospective client drops the prototype and breaks it. For these reasons, the durable material properties available with SLS can be highly beneficial.

Form, Fit and Function Review

Design analysis is perhaps the most important role of a rapid prototype. The ability to quickly verify a design prevents the investment of time and money in a poorly conceived project or component. Trade-offs between SLA, SLS and PolyJet are most obvious in this application. On one hand, SLA and PolyJet provide the tolerance and detail required for a form and fit review. On the other hand, SLS offers materials suitable for functional analysis. Not withstanding the strengths of SLA and PolyJet, SLS is most frequently recommended for this application. When a prototype will be subjected to mechanical or thermal loading, the material properties of DuraForm GF or Somos 201 are often critical to the success of the analysis.

Tooling Patterns

Rapid prototypes are extensively used as patterns for the creation of tooling. Common examples are silicone rubber, epoxy or plaster cast moulds. Tooling patterns are tolerant of a wide variety of material properties. This tolerance promotes the advantages of SLA and PolyJet. A tool and the part it produces can only be as good as the underlying pattern. Therefore, the surface finish, accuracy and detail replication available from these two technologies make them the preferred methods for pattern generation. Investment casting patterns require additional consideration beyond the requirements of other tooling patterns. Investment casting requires a pattern constructed in a material appropriate for burnout from within the ceramic shell. Both SLA and SLS have proven suitable for this application. However, some foundries prefer SLS CastForm while others prefer SLA patterns constructed with the QuickCast build style. Presently, PolyJet does not offer a solution for investment casting directly from a RP pattern.

Rapid Tooling

To further the advantages of RP, the technologies are being utilised to yield prototypes in production materials. In doing so, full analysis of a prototype design can be completed prior to the construction of production tooling. The term ‘Rapid Tooling’ has two meanings. Rapid tooling can be defined as the use of a rapid prototype as a tooling pattern, as described above, for the purposes of creating a tool for the moulding of production materials, or it can be defined as the direct production of a tool from a rapid prototyping system.

SLS has a distinct advantage in the direct production of tools. Designed for tooling applications, LaserForm delivers metal core and cavities with a robust composition of 60% stainless steel and 40% bronze. This technique utilises powdered metals coated with a binder. To process the part the binder is burned away, and the metal part that remains is infiltrated with bronze. Although many users have achieved significant time and cost savings with LaserForm tools, there are limitations when compared to traditional tooling methods. The most significant barriers are surface finish and accuracy. As discussed previously, SLS does not produce a part, or tool, with the precision available when machining. As a result, many users find that LaserForm must be machined to create a working tool that achieves production specifications. Some find that this additional work diminishes the value of this rapid tooling solution.

Conclusion

There is no single answer to which technology one should use. Likewise, the process selected will inevitably vary by project and perhaps by component. The final recommendation to be offered here is to consider the information provided and then further investigate the ability of the technologies in terms of the project itself

Article source: Time-Compression Technologies Europe magazine www.time-compression.com