Consolidation phenomena in laser and powder-bed …11].pdf · Consolidation phenomena in laser and powder-bed based layered manufacturing ... Selective Electron Beam Melting (EBM)

Annals of the CIRP Vol. 56/2/2007 -730- doi:10.1016/j.cirp.2007.10.004

Consolidation phenomena in laser and powder-bed based layered manufacturing J.-P. Kruth1 (1), G. Levy2 (1), F. Klocke3 (1), T.H.C. Childs 4 (1)

1 K.U.Leuven, Division PMA, Leuven, Belgium 2 FHS - University of Applied Sciences St. Gallen, Switzerland

3 Fraunhofer, Institute for Production Technology IPT, Aachen, Germany 4 University of Leeds, School of Mechanical Engineering, UK

Abstract Layered manufacturing (LM) is gaining ground for manufacturing prototypes (RP), tools (RT) and functional end products (RM). Laser and powder bed based manufacturing (i.e. selective laser sintering/melting or its variants) holds a special place within the variety of LM processes: no other LM techniques allow processing polymers, metals, ceramics as well as many types of composites. To do so, however, quite some different powder consolidation mechanisms are invoked: solid state sintering, liquid phase sintering, partial melting, full melting, chemical binding, etc. The paper describes which type of laser-induced consolidation can be applied to what type of material. It tries to understand the underlying physical mechanisms and the interaction with the material properties. The paper demonstrates that, although SLS/SLM can process polymers, metals, ceramics and composites, quite some limitations and problems cause the palette of applicable materials still to be limited. There is still a long way to go in tuning the processes and materials in order to enlarge the applicability of LM. This is not surprising if one compares it to the decades of R&D work devoted to tuning processes and materials for hot or cold forming, metal cutting (e.g. development of free machining steels), casting and injection moulding (including powder injection moulding: MIM, CIM, etc.).

Keywords Rapid Prototyping and Manufacturing, Selective Laser Sintering (SLS), Selective Laser Melting (SLM)

1 INTRODUCTIONLayered manufacturing (LM) goes back to the late 1980’s, early 1990’s [95, 98] with a clear breakthrough in 1994 at which time machine sales took off exponentially: see Figure 1 [216]. Today, distinction is made between Rapid Prototyping (RP) and Rapid Manufacturing (RM) [49, 121,164].

RP means the production of prototypes, visual design aids, touch, feel, fit and assembly test parts, etc., that are used in the product development phase and are not meant to be equivalent to real production parts at all levels. RM means the production of functional parts that are meant to be used as real production parts (end products) and should meet the various basic requirements put to such production parts. Rapid Tooling (RT) may be considered in this context as a sub-category of RM, i.e. production of functional tool components produced by layered manufacturing. [41,153, 181, 208]

Many layered manufacturing techniques exist today, the most popular being: photo-polymerisation (Stereolithography (SLA) and its derivates), ink-jet printing (IJP), 3D printing (3DP), Fused Deposition Modelling (FDM), Selective Laser Sintering or Melting (SLS/SLM and EBM 1 ) and to a lesser extend Laminated Object Manufacturing (LOM and similar sheet stacking processes) and laser cladding (LC) processes [98]. The importance of these technologies is confirmed by a recent study of NACFAM (National Council for Advanced Manufacturing, USA) that has identified Rapid Manufacturing as the most innovative and potentially disruptive manufacturing technology to emerge within the next 3 to 5 years [139]. Many of those techniques are however limited to Rapid Prototyping as they do not allow common engineering materials to be processed with sufficient mechanical properties (polymers, metals, ceramics, and composites thereof). In the prospect of Rapid Manufacturing, SLS/SLM seems to be the most versatile process, capable of 1

Selective Electron Beam Melting (EBM) may be considered as a variant of this category.

processing engineering polymers, metals, ceramics and a wide range of composites. In order to cover this wide range of materials, the laser processing of powder materials in SLS/SLM calls on various consolidation mechanisms: the binding of polymers, metals, ceramics and their mixtures or composites varies substantially with initial powder composition, the final aimed material composition and the aimed material structure (aimed porosity or density, microstructure, properties, etc.). Some process variants may apply post-processing after SLS/SLM: in such case, the selected layer consolidation mechanism should account for the post-process as well: e.g. need for open porosity if post-infiltration is applied.

Figure 1 : Yearly additive system unit sales worldwide [216]

This paper gives a survey of powder consolidation mechanisms applicable in SLS and SLM for various types of materials.

2 CLASSIFICATION

2.1 Types of powder consolidation mechanism (binding mechanism)

Laser-based consolidation of 3D parts from layers of powder material pre-deposited on a ‘build platform’ is commonly referred to as Selective Laser Sintering (SLS) or Selective Laser Melting (SLM). The distinction between SLS and SLM is rough, vague and does not cover all

-731-

types of consolidation. Various papers describe possible ways to consolidate material with a laser [18, 126]. A more detailed classification of consolidation mechanisms is given in Figure 2.Solid State Sintering (SSS) is a consolidation process occurring below the material’s melting temperature: diffusion of atoms in solid state (volume diffusion, grain boundary diffusion or surface diffusion [82, 156]) will create necks between adjacent powder particles that will grow with time (Figure 3) [66, 204]. This binding mechanism is only rarely applied in layer manufacturing (LM) as diffusion of atoms in solid state is slow and not very compatible with the desired high laser scan speed that should yield process productivity and economic feasibility. However, it is one of the consolidation mechanisms suited for consolidating ceramic powders [121]: see §2.2. SSS diffusion sintering is sometimes applied in post-processing of metallic or ceramic parts after initial (partial) laser consolidation using solid state, liquid phase or chemical binding [121]: see §4.

Figure 3 : Typical neck formed between two stainless steel powder grains obtained with Nd:YAG laser [204]

Liquid Phase sintering (LPS) and partial melting include a number of binding mechanisms in which part of the powder material is melted while other parts remain solid. The liquefied material will spread between the solid particles almost instantaneously as it is driven by intense capillary forces. This allows much higher laser scan velocities than for SSS. The material that melts may be different from the one remaining solid: the former ‘low melting point’ material is then often called the binder, while the ‘high melting point’ material is generally called the structural material. Different ways exist to bring those distinct binder and structural materials together:

mixture of two-component powder (i.e. separate binder and structural powder particles),

using composite powder particles that have a micro composite structure containing the structural and binder material, or

using coated particles where the binder material is applied as a coating around the structural material. The binder material (coating) will preferentially absorb the incident laser radiation. This enforces the intended melting of the binder.

When a different binder and structural material is applied, the two materials may differ substantially: e.g. polymer binder versus metallic structural material, metallic binder vs. ceramic structural material, low-melting metal (e.g. Cu) vs. high melting metal (e.g. steel or Fe). Examples and references are given in next sections. In all cases, the binder may be permanent (i.e. remain as part of the final product) or may be sacrificial (i.e. removed during a de-binding cycle). In both cases, the ‘green part’ coming out of the SLS machine might be porous (and become even more porous after de-binding) and might require a post-densification process: the latter being generally either a furnace infiltration process or a furnace post-sintering process (thermal sintering, HIP or other). In most cases, thermal post-sintering without applying pressure is not feasible: the porosity of the ‘green’ SLS part is still too high (25-50%) to initiate further densification by thermal heat only. Alternatively, this process may be used to produce porous structures [122, 168, 185].Partial melting is also possible even in cases where no clearly distinct binder and structural materials are used. SLS parameters are then adjusted to only partially melt the powder particles, which may be a single phase material or a mixture of different powders but without distinct binder powder material (particles): 1. When the heat supplied to a powder particle is

insufficient to melt the whole particle, only a shell at the grain border is melted. The core of the grain remains solid. This way the molten material will form necks between the particles and act as a binder between the nonmolten particle cores. This binding mechanism can arise as well with metals as with polymers, although the consolidation of polymer powders may also result from other mechanisms (consolidation at the glass transition temperature, which is lower than the melting temperature, polymer chain rearrangement and cross-linking).Such partial melting phenomenon and the resulting liquid neck formation were modeled at EPFL, Lausanne, [53]. Using a simple thermal model, skin and core temperatures of the powder particles were

Figure 2 : Laser-based powder consolidation mechanisms

-732-

calculated. This way the minimal pulse energy to fully melt the particle can be calculated. Below this value, the core temperature never exceeds the melting temperature and only partial melting is obtained.

2. Powders consisting of multiple phases or a mixture of different powder particles can be classified as partial melting when they are only partially molten. For example, the University of Leuven experimented with a Fe-Fe3P-Ni-Cu powder mixture, aiming at the production of fully dense parts [93]. The resulting part clearly contains unmolten Fe particles (see Figure 37).None of the 4 powders can be identified as a distinct binder material. The addition however of a melting point lowering additive like Fe3P or Cu3P is favorable in making the process more energy efficient: see details in §3.2.2, section B2.

3. Partial melting may also occur when using a single material powder having a bi-modal distribution: small particles are melted, while larger ones remains solid.

Liquid phase sintering and partial melting may have advantage over full melting (see below), even when aiming at full dense parts. The former allows the laser consolidation speed (scan speed, scan spacing…) to be increased drastically. The resulting economic benefit has to be considered against the disadvantage of the need for post-densification. Comparative studies demonstrated that molds made from PA-coated steel that are post-infiltrated with copper may not only be more cost effective [103], but also technically may perform better due to the beneficial properties of copper [106, 208].

Full melting is a third major consolidation mechanism, often applied to achieve fully dense parts without need for any post-process densification. Major progress has been achieved towards full melting of metal powders up to densities of 99.9% by applying modern laser sources and optics yielding high energy densities in the spot [105].Dedicated SLM machines are now offered by 4 German vendors [58, 103]. As compared to early day SLS machines equipped with CO2 lasers, these machines apply solid state lasers: diode pumped Nd:YAG, fibre or disc lasers [150].Full melting has the main advantage to produce almost full dense products in one step, but also has drawbacks that require careful process control: The high temperature gradients and densification ratio

during the process yield high internal stresses or partdistortion (from 50% powder porosity to 100% density in one step) [6, 134].

The risk of balling and dross formation in the melt pool 2 may result in bad surface finish.

Chemical induced binding is a fourth main consolidation mechanism. Today it is not commonly used in commercial LM equipment, but it turns out to be a feasible consolidation mechanism for polymers, metals and ceramics. IPT-Aachen has applied chemical binding for laser sintering SiC ceramic powder [88]. Chemical induced binding of metals could e.g. be invoked for sintering Al powder by making the Al react with the N2 atmosphere used commonly in SLS machines, thus creating an AlN binder phase holding the Al particles together. Such mechanism has been used by the University of Queensland, be it during post-sintering in a N2 furnace, rather than during the initial SLS phase itself, at which stage a sacrificial polymer binder was used to consolidate

2 Those melt pool phenomena have been studied within the CIRP STC-E Working Group on “Melt pool phenomena in SLS/SLM/LC” that met in Leuven (Jan. 2004), Aachen (Oct. 2004) and Leeds (Sept. 2005). They are dealt with in §2.4 and 3.5.

the green Al part [172]. Other (potential) examples of “Selective Laser Reactive Sintering” were investigated at the University of Texas at Austin [11]. Some investigators tried to apply self-propagating high-temperature synthesis (SHS) in which a low laser power induces a self-propagating exothermal reaction capable of self-heating and consolidating the powder in a controllable way [107,177,182]. This allows reducing the laser power significantly. The metallic liquid phase stays longer and the accuracy of the produced piece will be higher. Shi et. al. applied a SHS process to laser sinter metallic materials (Ni-Al, Al-Ti, Ni-Ti) [176]. The Osaka Sangyo University tried to produce cermets (Al2O3-Cu), ceramics (MoSi2),and intermetallics (TiAl) by SLS induced chemical reaction [80]. This will be discussed in §3.4.7.

2.2 Type of consolidation mechanisms versus materials

The applicable type of consolidation is basically dependent on the material being processed. Table 1 gives an overview of what material can be basically consolidated with the various binding mechanisms.

2.3 Influence of laser type Different types of lasers are used for powder-bed based RM: CO2, Nd:YAG (lamp pumped, diode pumped, or Q-swithed) [60], fibre lasers, disc lasers, Cu-vapour lasers [76]. The type of laser has a large influence on the consolidation of powder particles, because: laser absorption of various materials greatly depends

on the laser wavelength: e.g. high absorption of polymers and oxy-ceramics at 10.6 μm wavelength of CO2 laser; high absorption of metals and carbo-ceramics at ±1 μm wavelength of Nd:YAG, fibre and disc lasers [105, 196];

the possible consolidation mechanism highly depends on features like energy density: e.g. SLM of metals and ceramics requires a high energy density that is more easily achieved with fibre and disc lasers (high beam quality, i.e. high focusability);

laser mode (continous, pulsed, Q-switched,…) also has a large influence on consolidation (see end of §2.4).

Several researchers investigated the influence of different laser types on the absorption and consolidation of powders in SLS/SLM. Some did it experimentally [100,114, 156], others by simulation [114].

2.4 Problems in consolidation There are two main general problems in consolidation. One, applicable to all process variants other than the full melting of metals, stems from the short material heating times caused by the scanning laser beam, relative to the time required for consolidation, and the fact that it is only temperature effects and gravity and capillary forces that can provide the driving force (there is no mechanical pressure, as in moulding processes). It can lead, as has already been written, to porosity in parts. Post-processing is then required if a pore-free material is needed. For polymer processing, it means that melt viscosity must be kept less than some critical value. For example Nylon-12 (PA 12) with an in-process viscosity 100 Pas can be fully densified under surface tension driving forces 30 mN/m ( / 0.3 10-3 m/s), but polycarbonate (viscosity 5000 Pas, and similar surface tension) can not [28].The other problem, important for the full melting of metals, is quite the opposite. Here the viscosity is about 1 to 5 mPas [16], the surface tension about 1 to 2 N/m [12] and consequently / 100 to 1000 m/s. The problem is not time for flow, but controlling the flow [87]. The pool of

-733-

molten metal must wet the previously processed metal below it (except on the rare and sometimes problematic occasions that it is an overhang). And when it solidifies, its upper surface must be flat enough to enable a next layer of powder to be spread over it. Both materials properties and processing variables can influence these things. As far as wetting the material below is concerned, contamination, usually by oxygen, and in that case either as oxide, impurities within the powder or as trapped gas in the powder bed, is a main materials cause of consolidation problems. For example, the surface tension of pure liquid iron near its melting point is almost 2 N/m, but 0.02 wt.% of oxygen reduces this towards 1 N/m [12,110]. Liquid iron will not spread out over an iron oxide covered surface. Even in the absence of such contamination, there may be a problem of a liquid metal wetting its solid form, if the solid has almost the same temperature. In this homologous wetting case, there is no driving force for wetting. It has been suggested on the basis of experimental studies on a stainless steel [35] that it is necessary for the previously processed material to be re-melted by the currently melted material. The excess energy (and hence the higher peak temperatures in the melt pool) required for this will have consequences for the quality of the upper surface of the melt pool. When temperature gradients are created in a melt pool, there is the potential for convective motions in the pool to reduce those gradients. Temperature gradients through the depth of the pool, causing density gradients, and hence buoyancy forces, are probably of no importance to the millimeter or less melt pool sizes of SLM (compared to the much larger sizes that can make buoyancy effects relevant to welding [137]), although see [142]. But temperature gradients in the surface, coupled with a temperature dependent surface tension, can cause rapid motions (tens of cm/s or more), known as thermo-capillary or Marangoni flow. The equation below defines the dimensionless Marangoni number Ma.

ad w d dT wMdx w dT dx w

(1)

It is the ratio of the speed by which a surface temperature gradient dT/dx may be reduced by convection or conduction (w is the linear size of the pool, is its thermal diffusivity) [137]. For typical ferrous material property values, also supposing an initial temperature gradient dT/dx to be due to the difference between liquidus and solidus temperatures over the half-width of the melt pool, and w = 1 mm, Ma can be calculated to be of the order 1000. Marangoni flow dominates. When d /dT is negative, as it naturally is with decreasing surface energy from the liquidus to the vapourisation temperature, the Marangoni flow is from high to low temperature, that is to say from the centre to the edge of the melt pool, and a flattened or dished surface is created.

3HM = hardmetals

But with oxygen contamination of steels, and dependent on sulphur and nitrogen content too, d /dT can be positive [83, 159]. Then the melt pool becomes humped. Again, such effects are clearly seen in welding [127, 128], as well as in SLM [156, 158].In the previous paragraphs, different effects at the under and top side of the melt pool have been considered. But in SLM there is a leading edge to the melt pool too, where the surface, subject to the surface tension induced flow, is advancing into powder that at some distance ahead of the melt front will certainly still be below the solidus temperature. The wetting or not of this solid powder and, if it is wetted, its dragging or not into the melt pool, are further complications of the consolidation process that have not been systematically studied. It may be that the compromises needed to balance good wetting or remelting of previously melted material, with the top surface of the melt pool freezing sufficiently flat, and the powder bed not being disturbed too much ahead of and around a currently melting track, will put a large constraint on metals and powder preparations that are suitable for processing by SLM. Much basic work remains to be done in this area. The consolidation flows and the materials dependent phenomena accompanying them and discussed above differ from flow instabilities in a pool that can arise because of its shape. Long thin melt pools are known to break up into balls, called “balling” and commonly described as due to Rayleigh instabilities [3, 19, 74, 87,135, 136, 157]. For a free melt pool, as would describe a single melted track in a deep powder bed, balling is predicted for pools (imagined to be cylindrical) of length to diameter ratio greater than . Balling as this ratio is approached is reported by [27, 156, 157]. That work also showed length to diameter ratio increasing with laser scan speed. Thus in this case, instability limits processing to a low scan speed. A melt pool on a solid substrate can also be unstable, depending on its contact behaviour with the solid [38, 169]. A limiting length to diameter ratio of 2.1 has been reported [96], also creating an upper limit to the laser scan speed. Most balling studies have been performed in conditions of single line or raster scanning powder beds when the laser-caused temperature rise has essentially been due to the current line scan. At higher scanning velocities and/or shorter vector lengths, however, previously scanned tracks are not yet solidified when a new track is being scanned. Thus, the actual width of the melt pool is enlarged, since it now spans multiple scan tracks [135,136]. Therefore, the length-to-width ratio may actually be reduced (even if the melt pool length enlarges), resulting in a more stable behaviour compared with a single track scanned with the same parameters. Mercelis [135]showed that melt pool dimensions may depend much more on the geometry of the 3D part and the 2D scanning path, than on the laser power and scanning velocity. Therefore, also the scan track instability cannot be

Solid State Sintering Liquid Phase Sintering Partial melting Full melting Chemical

Polymers No Yes (e.g. PS) Yes (e.g. PA) Seldom (e.g. partial cross linking PMMA)

Metals Seldom Yes (many kinds) Yes (steel, Ti...) Yes (e.g. Alu)

Cermets/HM 3 No Yes (e.g. WC-CO) No Yes (e.g. Al2O3 – Cu)

Ceramics Yes (Phenix) Yes (e.g. SiC) Yes (e.g. ZrO2) Yes (e.g. SiC)

Other composites No Yes No Yes

Table 1: Consolidation mechanisms versus materials

-734-

examined without taking into account the scanned geometry. That brings this section finally to the influence of process variables on problems in consolidation. First is the influence of laser type (see also §2.3). The higher the laser absorption into the powder bed, the less scope there is for variation in absorption, and the easier is the process to control. Pulsed laser beams and smaller beam diameters may lead to smaller, more stable melt pools, and in some cases pulsing may result in plasma formation and recoil forces adding a mechanical component to consolidation phenomena [97, 144, 157]. Smaller layer thickness reduces the need for over-heat if re-melting the previous layer is important. Thinner layers means finer powder sizes. These in turn are easier to melt due to their higher surface-to-volume ratio. Processing changes that lead to smaller melt pools also seem to lead to easier consolidation. It may be that a final problem in consolidation is that it cannot be scaled up to faster processing speeds.

3 DISCUSSION OF MATERIAL CLASSES In this section the consolidation phenomena are discussed per material class:

Polymers: i.e. substance with a high molecular weight which is made of many repeating smaller chemical units or molecules. Metals: i.e. substance whose atoms are connected by metal bonds. Ceramics: inorganic and non-metallic materials. Cermets and hard metals: i.e. composites in which ceramic particles are bound in a metal matrix (binder). Composites: mixtures of different materials which result in an inhomogeneous compound. Cermets and hard metals may be considered a special class of composites.

3.1 Polymers Even though polymer is the most processed type of material in SLS/SLM, the consolidation phenomena invoked for polymers are probably still amongst the least understood or at least the least described in literature [81].This explains why commercial applications of SLS today are limited to a small number of polymers: mainly polyamide (PA 12 and PA 11), and some polycarbonate (PC) [48], polystyrene (PS) and variants of those [175].

Figure 4: Thermoplastic polymers (Red = material used in SLS)

In literature there is no agreement on the nomenclature of the physical state of parts after laser powder processing. Laser consolidation of polymers normally involves meltingof thermoplastics (partial-SLS or full-SLM). A clear distinction should be made between (Figure 4):

(semi-)crystalline thermoplastics amorphous thermoplastics.

When heated up from very low to very high temperatures, all those thermoplastic materials will change from a hard (solid and glassy) structure to a softer (tough leathery or rubbery, solid or non-pourable) structure and finally turn into a viscous flowing melt. Materials however differ in the way and the temperatures at which those transitions occur. Semi-crystalline polymers have a glass transition temperature Tg that is below or around room temperature (-100° to 50°C) and a distinct melting temperature Tm that is above 100°C (between 100° and 400°C) at which a significant volume change happens. Amorphous polymers do not depict a clear melting temperature range. They have a glass transition temperature Tg that lies around 100°C and above which the material will gradually evolve to a leathery, rubbery and finally liquid state as temperature increases, without clear transitions. Figure 5indicates the glass transition range ( Tg) and melting range ( Tm) for three important semi-crystalline polymers (PE, PP, PA 6) and indicates for three amorphous polymers (PS, PC, PMMA) the glass transition temperature ranges ( Tg) and the temperature range in which the transition to leathery material goes quickly ( Tf).This last range ( Tf) can physically not be defined exactly.

Figure 5 : Phases and transition temperatures of some polymers

Notice also that the Tg and Tm values largely depend on the molecular weight (MW) of the polymer. This explains why there might be a significant difference in SLS processability between a low and a high molecular weight PA or PE. The temperature transitions and melting range of polymers can be observed by recording a DSC plot by Differential Scanning Calorimetry analysis. It measures the difference in the amount of heat required to increase the temperature of a sample and a reference as a function of temperature. Both the sample and reference are maintained at very nearly the same temperature throughout the experiment. Figure 6 gives the DSC plots of Duraform PA 12 and PA 6 [166]. The plots illustrate the much smaller melting range (small width of peak around 187°C) and the smaller melting temperature (Tm=187°C) of PA 12 as compared to PA 6 (Tm=223°C). This explains, besides viscosity differences (see §2.4 and §3.1.1), why PA 12 is more prone to SLS than PA 6 (or PA 66 having Tm= 262°C and a wider peak than PA 12) [212].The laser consolidation of (semi-)crystalline polymer powder will happen by heating above their melting temperature Tm. Semi-crystalline materials have a highly ordered molecular structure with sharp melt points. They do not gradually soften with a temperature increase but rather remain hard until a given quantity of heat is absorbed and then rapidly change into a viscous liquid.

-735-

The molten polymer flows in between the powder particles, forming sintering necks. With enough heat, the complete layer is fully molten and overlaps to the previous layer. As the molten polymer cools down below Tm,polymer crystals nucleate and grow, recreating regions of ordered molecular chains (crystallites) mixed up with disordered amorphous regions. Above Tm, the polymer depicts a relatively low viscosity, as compared to amorphous polymers, which favours the rate and amount of consolidation. The densities obtained will be close to full density and the mechanical properties will be close to those of moulded polymers. However, the freezing of the polymer at Tm coincides with an important shrinkage (phenomenon not occurring with amorphous polymers), that may induce geometrical inaccuracies and distortion of the part (Figure 13). A good way to prevent this is to preheat the polymer powder to a temperature slightly below its melting temperature and keep it there for a certain time after consolidation [26,197, 212].

Figure 6 : DSC plot of PA 12 (Duraform) and PA 6

The consolidation of amorphous polymer powder

r semi-crystalline polymers we have rather

thermal behaviour of polymers is amplified

occurs by laser heating above the glass transition temperature, at which the polymer is in a much more viscous state than semi-crystalline polymers at similar temperature. Amorphous polymers have a randomly ordered molecular structure. They do not have a sharp melt point but instead soften gradually as the temperature rises. The viscosity of these materials changes when heated, but they seldom are as easy flowing as semi-crystalline materials. The flow and sintering rate will be less, resulting in a lower degree of consolidation, higher porosity, less strength, but also a lower shrinkage that is favourable in cases of SLS of patterns for producing mould for moulding or casting (i.e. Indirect Rapid Tooling, investment casting patterns, etc.). In several applications post-infiltration of the pores is applied to consolidate the part (see § 4). In conclusion fofull melting consolidations whereas for amorphous polymers we rather see a partial melting binding mechanism.This complexin SLS processing due to several factors. The costly pulverization either by cryogenic milling or precipitation-process (Figure 7) limits the availability of polymer powders. The pulverization can influence also the

consolidation process: the powder layer’s thermal characteristic differs.

The material is in powder form with a certain particle size distribution (Figure 8) and the induced energy, heating the fine powder particles, may lead to evaporation or disintegration which in turn disturbs the process. It shields off the laser window and the IR heating elements. In short, this has a negative influence on the process consistency and the economy of the process. Further the fine particles may create either a dusty atmosphere or coagulations clustering preventing the homogeneous deposition of layers, thus preventing good layer consolidation.

a

b

Figure 7 : Powder particle shapes REM, a) precipitation, b) Cryogenic milling

A variety of commercially available technical polymers with excellent properties for SLS have been tuned by blending or by special design of the chain length (analogy to molecular weight ranges). Often they result in a multi-peak DSC diagram. SLS can handle only one fixed controlled consolidation temperature or several very close peaks (typically 5-10 °C difference), which means that polymers with more than one tight melting range are not processable with SLS. Figure 9 demonstrates the DSC curve of a commercial pulverized blend with no real full melting SLS processability, due to two different melting points at 178°C and 185°C. We can associate seven main SLS polymer powder materials with different application fields. These material families can also be attributed to a specific consolidation mechanism as discussed in §2.The four main thermoplastic SLS applications and the related thermoplastic material are:

a. SLS of polymer parts: semi-crystalline polymers (§3.1.1)

b. SLS of investment casting patterns: amorphous polymers (§3.1.2)

c. SLS of metal or ceramic parts using a sacrificial polymer binders: thermally degradable amorphous polymers (§3.1.3)

-736-

d. SLS of reinforced or filled semi-crystalline

eric material parts (§3.1.5) in the racing

mosetting materials (§3.1.7)T

polymers for highly loaded parts (§3.1.4)SLS today also allows production of parts using:

e. Flexible elastomf. Polymer-polymer blends (used

industry) (§3.1.6)g. Ther

hose different material categories are discussed in more detail below.

Figure 8 : Typical particle size distribution of Duraform PA 12 (Source FHSG St. Gallen)

Figure 9 : Commercial pulverized blend with no real ltcurve (right)

[151,

on is then obtained by keeping the powder bed

ay be raised to compensate for that, but only in a narrow range and with limited effectiveness (see also § 3.1.8).

a

SLS processability: (left) merecrystallisation curve

3.1.1 Semi-crystalline polymers Today, the production of prototypes (RP) and functional parts (RM) by SLS is basically limited to nylon, i.e. polyamide (PA) [7, 22, 24, 26, 84, 102, 218]. Quite some research is going on using other (semi-)crystalline thermoplastics: polyethylene (PE) [155, 165], PEEK152], PCL [148, 203, 214]. and PEEK are of special interest as biocompatible polymers [151, 192, 214].The fundamental reason why PA sinters well, while more difficulties are experienced with other polymers is not totally clear yet. Some semi-crystalline polymers are able to crystallize much faster than others, as a result of their chain structure. The rate of crystallization is minimum near Tg and Tm, and reaches a maximum between those two temperatures. Crystallization rate should however be kept relatively slow (relative to the vertical build rate) in order to avoid part distortion due to freezing shrinkage. Good crystallizatiat high temperature for a sufficient long time after sintering.To obtain fully dense parts, the melt viscosity should be low enough to allow complete consolidation within the time scale of the process. Pure polymers with melt viscosity in the range of few tens to few thousands poise (e.g. nylon and waxes) can be processed successfully and reach near full density [11]. However, even for PA-12, the SLS part still contains small amounts of porosity and the

process is something between partial and full melting. The densities obtained in SLS of PA 12 are just slightly below those of compression moulded parts (0.95-1.00 versus 1.04 g/cm3), yielding very similar mechanical propertiesunder compression, but somewhat lower tensile strength and notched Izod values that are sensitive to small voids. The melt viscosity is linearly related to the molecular weight (MW). For good interfusion of the polymer chains, sufficient to provide particle necking and layer to layer adhesion, a low melt viscosity is desirable. However, low viscosity results in high shrinkage and poor part accuracy. Thus an optimum MW range exists but is not easily controllable, neither at the polymer production nor during aging in process. The thermal degradation over time when re-using the same powder repeatedly, results in the decay of the MFI (Mold Flow Index) hence a drop in powder flowability and a rise in melt viscosity preventing proper sintering with reasonable quality: the surface gets an unsmooth “orange peel” texture (Figure 10b) after a certain number of runs (Figure 10a). A drop in mechanical properties can be observed as well. The consolidation temperature m

b

Figure 10 : a) The rise in viscosity of re-used PA 12 material, b) resulting orange peel texture on SLS part

To achieve functionally strong prototypes, all powder bed based processes need to work with polymers that have molecular weights as high as possible [155]. High molecular weight can lead to difficulty in placing and forming the material because of the high viscosities involved. Increasing the molecular weight of the polymer

oxidation resulting in de-polymerization). Typically

powder can reduce shrinkage in the sintered part and thus improve the dimensional accuracy. PA and other semi-crystalline polymer powders for SLS are normally supplied without additives. Some additives may be added to favour the powder fluidity and powder spreading, but those have no influence on viscosity and sinterability. Atmospheric control in the SLS chamber is very critical in terms of temperature (control of sintering and shrinkage), humidity (powder fluidity) and oxygen content (to avoid thermal degradation of the polymer by

Cooling curve

Heating curve

Orange peel surface texture

-737-

polymers like PE and PP are more prone to rapid oxidation than aromatic backbone polymers like PET

e to the purity differences compared to the first powder.

a

(semi-crystalline) or PC (amorphous). Several researchers investigated SLS of biocompatible semi-crystalline thermoplastics like PEEK and PCL (Poly-ε-caprolactone) [9, 31, 148, 152, 170, 192, 214].PCL is a biodegradable linear polyester, while PEEK is not biodegradable. A polymer has never one specific molecular weight, but consists of a mixture of molecules with different chain lengths, and so, different molecular weights. The mean molecular weight and the molecular weight distribution are important properties of a polymer. Typically SLS of PCL with MW of 50.000 g/mol demonstrated good laser sinterability [214], while PCL with a MW of 40.000 g/mol could not be laser sintered well [203] due to big distortions caused by shrinkage. This shrinkage is due to the lower molecular weight of the second PCL powder, and du

b

Figure 11 : a) PA 12 Powder, b) tensile break cross section showing some air voids (Source FHSG-St.

Gallen)

break cross section

oulded parts concerning

makes them well suited for investment casting patterns.

a

Powder particle morphology also influences the laser consolidation. Figure 11a shows powder (PA 12) with a near to spherical shape whereas Figure 12a shows irregular shaped PS powder. When no pressure is applied during consolidation, voids and air traps remain in the solid part. This reduces the effective cross section and the tensile strength compared to moulded parts from the same polymer (10-20% reduction). Figure 11b (PA 12) and Figure 12b (PS) depict the tensiledemonstrating partial melting regime.

Comparison between SLS and mstrength is discussed in §3.1.8 .

3.1.2 Amorphous polymers Amorphous polymers with strongly temperature-dependent viscosities (or high activation energy for viscous flow) have been readily processed using SLS [11,14, 75, 141]. Typical examples are polycarbonate (PC) and polystyrene (PS). Depending on the powder size and MW, they are generally pre-heated to a temperature near or even above the glass transition temperature Tg and will

melt as the temperature further rises above Tg. They mainly have the advantage that their volumetric shrinkage, when cooling down, does not show a jump, as it does for semi-crystalline polymers: see Figure 13. This fact provides high accuracy with little distortion. However, amorphous polymers show a lower level of consolidation (higher residual porosity) and hence do not reach the strength of their moulded equivalent. The higher degree of porosity, low thermal shrinkage/expansion and high accuracy

b

gure 12 : a) PS Powder (CastForm), b) Tensile breaFi kcross section demonstrating partial melting regime

us polymers for 3.1.3 Thermally degradable amorphoSLS of metal or ceramic parts

Various SLS applications make use of a powder combining a degradable polymer binder and a structural material (metal or ceramic) that produces a “green part” in which the structural particles are bound into a degradable polymer matrix [77]. These ‘green parts’ are further processed in a furnace to burn out the polymer (“de-binding”) and to consolidate the remaining polymer-free metal or ceramic part by post-sintering, infiltration, or HIPing: see § 4. The polymers used as the sacrificial binder in those applications are different from the polymers described in the previous and subsequent sections. They should allow de-binding, i.e. thermal depolymerization. Polymers used for this purpose are PMMA and copolymers MMA-BMA that can be de-binded in a furnace at temperatures around 350-450°C. They should not be water soluble [11, p. 112] and should have sufficient strength, even when heated up in the de-binding furnace, in order to avoid that the ‘green part’ would collapse before some solid state sintering can consolidate the metallic or ceramic particles together. The polymer must also be fully de-bindable (i.e. depolymerised and sublimated) and should not yield excessive contamination of the remaining part with carbon residuals (important e.g. for steel parts). If the strength of the polymer in the green stage is insufficient, the green part can be consolidated by cross-linking the binder. This can be done by impregnating the green part with a water soluble thermosetting acrylic emulsion and drying it in an oven at 50°C, before it goes to the de-binding furnace [11, p. 139]. Alternatively, a cross-linker can be added into the degradable polymer

-738-

coating of the powder [11, p. 139]. This will eliminate the step of impregnating the green part. When using sacrificial binders, care should be taken because depolymerization could also occur during SLS [11 p. 108, 82, 88]. Indeed, these polymers are deliberately designed to depolymerise as an aid to their thermal removal in post-SLS processing. Polymers based on copolymers of n-butyl methacrylate (BMA) and methyl methacrylate (MMA) have been observed to also depolymerise during SLS processing [200, 201].

Figure 13 : Comparison of relative volume between amorphous and semicrystalline polymers [11]

a

b

cFigure 14 : SLS of AW glass ceramic with sacrificial

binder (a) Powder mixture (b) Green part (c) Brown part

The polymer binder may be added to the structural material in two ways: mixing [34, 123] or coating [171,172, 199, 217]. Typical volume of binder in the powder is 5% for mixed powders [34] as well as coated ones [11,

rates three stages in the SLS

e just enough to give the

still is not totally dense and may need further densification to convert the brown part into a final one.

p136]. The latter typically corresponds to a 5 μm polymer coating on a 55 μm 1080 carbon steel grain [200].Figure 14 [108] demonstprocess of Apatite-Wollastonite (A-W) glass ceramic with 5 wt% MMA-BMA binder: a) powder mixture before processing (white particles are

A-W glass ceramic, grey particles are MMA-BMA), b) highly porous part after SLS processing (The few

polymer connections ar'green' part sufficient strength to be transferred to the post-treatment furnace)

c) 'brown' part after polymer debinding and 1 hour firing at 1150°C. During firing, A-W is partly melted. The part

Figure 15 : Glass-nylon powder mixture (Source: K.U.Leuven)

Figure 16 : SEM micrograph of HA particle distribution in HDPE matrix [165]

Figure 17: PA-12 GF One of the most popular glass beareinforced HTD materials (Source FHSG St. Gallen)

ds

binders described in

3.1.4 Semi-crystalline reinforced polymer for metal, ceramic or glass reinforced plastic parts

Today there exist several SLS powders aimed at production of reinforced polymer parts. Applications include parts made from glass reinforced PA [28, 30], Cu filled PA [20], Al filled PA [1, 46], SiC-PA [59, 76], HA-HDPE, i.e. hydroxyapatite reinforced high density polyethylene [165], HA-PA [69, 166], Polysiloxane-SiC [54]. Unlike the sacrificial polymer

-739-

§3.1.3, the polymers used here should withstand thermal degradation and should be durable. The initial powder may be a mixture of polymer particles and reinforcement beads. Figure 15 shows such a mixture of spherical PA and glass particles: on the right side of the picture, the glass particles are colored blue with Prussic ink for better visibility. Alternatively, the single powder particles may already be a composite consisting of a polymer matrix (PA or PE) containing filler particles [165,166] (see particles[20].

Figure 18 : New PA powder with a high-aspect ratio filler

a

Figure 16) or a polymer coated filler

(Source FGHS St.-Gallen)

bFigure 19 : a) PA 12 and 30% Al blend; b) a tensile break

turned out to be better

heat concentration on the filler has a negative influence on

only (for PA 6-8 times depending on judgment

s, strength

ls for Shore A hardness and elongation at break.

a

cross section with defects due to inhomogeneity (Source FHSG St. Gallen)

Figure 17 shows the fracture surface of a part made with a popular commercial glass filled nylon powder. One can see that the connection between the glass beads and PA binder was not perfect, that many glass beads have been pulled out and that the PA was stretched during breakage, resulting in a rough spongy surface. Figure 18 shows a new reinforced SLS powder having elongated filler beads (patent pending). This new powder that the traditional GF-PA powder: tensile strength x 1.8; E-modulus x 1.3; elongation x 3.3. Aluminium filled PA 12 often shows coagulation, since mixing the aluminium and PA powder particles isn’t always very succesfull. This causes segregation during layering due to the differences in size and specific weight (Figure 19). This fact is even more striking in PA 12-Cu. The inhomogeneous solid with voids creates mechanical defects and deteriorates the tensile strength. The local

the binder polymer causing a very short recycle ability of 2-3 timescriteria).

3.1.5 Elastomeric materials Elastomeric polymers have a structure with long chains and only few cross-links between them. Below the glass-transition temperature they are brittle, above they are very elastic. When the density of cross-links enlargeand stiffness also enlarge, and ductility lowers. A new elastomer powder material for the SLS process was recently released [120]. Figure 20 demonstrates this polyester based elastomer (patent pending). The most significant properties in the thermoplastic material selection were considered and compared with reference to elastomeric materials as Neoprene, EPDM and natural rubber. Figure 21 gives a comparison for the different elastomeric materia

bgure 20 : Polyester based elastomer a) Green par

r sintering at low hardness b) Infiltrated wFi t

afte ith polyurethane (Source FHSG St. Gallen)

relation to commonly HSG St. Gallen)

Figure 21 : Developed elastomeric material for SLS in used elastomers (Source F

0

25

50

75

100

0 1 2 3 4 5 6 7 8Material

Shor

e A

Har

dnes

s

0

250

500

750

1000

Elon

gatio

n %

Shore A (low) Shore A (high) Elongation %

-740-

Sintaflex not infiltrated

15W13W11W5W

9W7W

0

20

40

60

80

100

100 150 200 250 300 350Elongation (%)

Shor

e A

Har

dnes

s

Sintaflex infiltrated

5W 7W 9W 11W 13W 15W

0

20

40

60

80

100

100 150 200 250 300 350

Elongation (%)

Shor

e A

Har

dnes

s

Figure 22 : Hardness Shore A vs. Elongation [%] for infiltrated and not infiltrated parts processed with different laser powers (Source FHSG St. Gallen)

a

Sintaflex: E module vs. LS Laser Power

0

5

10

15

0 5 10 15 20Laser Power [W]

E M

odul

us [M

Pa]

not infiltratedinfiltrated

b

Sintaflex: Rupture at break Rb vs. E Module

0

2

4

6

0 2 4 6 8 10 12E Modulus [MPa]

Rb

[Mpa

]

RbRb (inf.)

Figure 23 : (a) The elastic modulus [MPa] against laser power for infiltrated and not infiltrated parts

(b) The tensile strength Rb [MPa] at ruptures against elastic modulus (Source FHSG St. Gallen)

The influence of laser processing parameters on Shore A hardness and ductility is shown by Figure 22. Clearly the parts processed with higher laser power become more ductile and have a higher Shore A hardness. The influence of laser processing parameters on elastic modulus E and the tensile strength Rb at rupture is shown by Figure 23.

3.1.6 Polymer blends Polymeric blends offer an alternative mean to obtain SLS parts with specific structure and properties, permitting the development of new applications [174]. Most polymeric

blends are multiphase systems and, therefore, their properties largely depend on their microstructure [162,215]. Salmoria, for instance, used blends of PA and HDPE (blend ratio’s of 80/20, 50/50 and 20/80 wt%) to achieve dedicated properties. Depending on the blend ratio, different phases and micro-structures were observed using SEM, EDX and XRD analysis [162].FH St.Gallen experimented with powder blends (PA12 with additions of PA11 or Polyester) mainly to enhance the elongation at break. Recently, a company from Austin started commercializing a mixture of PA12 and PA11 that demonstrates two DSC melt peaks at 185°C and 187°C respectively. The process is set to only melt the lowest peak. This clearly suggests a liquid phase sintering consolidation process.In a US patent [40] the following is claimed: A particle for use in selective laser sintering (SLS), including a core (1) formed from at least one first material, and at least partial coating (2) of the core (1) with a second material (further components are optional), the second material having a lower softening point than the first material, wherein the softening point of the second material is lower than approximately 70°C. The coating (2) generally contains a polymer, preferably a thermoplastic polymer, e.g. a polyvinyl acetal, preferably a polyvinyl butyral. It may consist of alloys with a low softening point which are used e.g. in fuses. Moreover saturated linear carboxylic acids with a chain length of ≥16(e.g. heptadecanoic acid, melting point 60-63°C.) or polymers in the broadest sense may also be suitable. The softening point of the 2nd material of approximately 70°C or below, allows laser sintering to be carried out at significantly lower temperatures compared to particles which have been used hitherto, and therefore also allows a significantly lower temperature difference between irradiated particles and standard room temperature. Tests have shown that the lower maximum temperature difference also improves the temperature homogeneity of the building space as a whole.

3.1.7 Thermosetting materials Thermosetting polymers can be used in different steps of the SLS process. These materials can be used as an infiltrant [47], because it is a relatively inexpensive path to a fully dense, stiff, net shape, polymer matrix composite part. In addition this is a very useful intermediate step for fully functional materials, as critical surfaces and tolerances may be achieved easily at this stage. An example of infiltration with thermosetting polymers is the production of metal-epoxy molds via SLS indirect processing [11, p143]. In this process, SLS is used to form green mold cavity inserts from metal powder that is coated with fusible thermoplastic binder. In subsequent steps, the binder is thermally removed and the metal powder is oxidized to form a porous metal/ceramic cavity that shows little shrinkage and generally excellent retention of geometry, relative to the green part. The cavity is then strengthened and sealed by infiltration and cure of an epoxy tooling resin. It is also possible to produce metal parts by laser sintering a mixture of metals with thermosetting polymers [123].When a thermosetting material is exposed by a laser source, the thermosetting material turns into a viscous liquid instantly. With for instance SLS of epoxy resin mixture with iron powder, the polar groups (e.g. epoxy group) in the molecule of the resin are activated simultaneously. The liquid flows penetrate the ‘pipes’ made from pores and wet metal particles. As such, they make bridges from one particle to another. The binding effect, depicted in Figure 24, is dominated by the interfacial characteristic between the resin and the iron.

-741-

Iron surfaces commonly attach some active hydrogen atoms because of its attraction of some molecule such as H2O and HCl due to its high polarity. Hydrogen bonds occur between the electronegative oxygen atoms in polar groups in resin molecules and the electropositive active hydrogen ones on iron surfaces. Iron particles are strongly bonded because hydrogen bond attraction is more intensive than that of inter-molecular action existing on the interfaces between iron surfaces and other nonpolar polymers. The resin viscosity is lowered at higher temperature (i.e., laser energy) and its viscous liquid can spread easily. Thus many more iron particles surfaces can be attached by resin. However, degradation of resin, induced by excessive laser energy, might occur and reduce the bonding capability.

Figure 24 : Binding mechanism for epoxy resin to iron particles

Injection vs. SLS Materials

0

4000

8000

12000

16000

0,0 100,0 200,0 300,0 400,0Elongation [%]

Tens

ile m

odul

us [M

Pa] Injection SLS

c

23

45 6

8 9

ab

d e f

1

7

0

50

100

150

200

250

0,0 50,0 100,0 150,0 200,0 250,0 300,0 350,0 400,0

Elongation [%]

Tens

ile s

tren

gth

[MPa

] Injection SLS1

2

3

4

10

67

8

9

a

c d e f

Figure 25 : Commercial available SLS materials (yellow) in comparison to some injection polymer materials (in blue)

3.1.8 Properties of molded and SLS polymers Although great efforts have been made and some results were achieved, the reasons for the relative few material options are many. Some are surely of interest to this paper about consolidation, however when dealing with polymers other phenomena have to be considered as to enhance the binding models.

By the end of the day we want to have nearly the choice as e.g. in injection moulding. The obvious way is to compare the differences and compensate for deficits. Figure 25 shows all the commercial available SLS materials (in yellow) in comparison to some injection polymer materials (in blue): the comparison is shown for tensile strength, tensile modulus and elongation. We find some deficits. While some SLS materials have comparable properties to molded ones, none of them is able to reach the highest values of tensile modulus and strength, nor elongation. The low elongation capability can be explained by the many short binding necks creating high stiffness with little elongation. A further interesting finding indicated by Zarringhalam [218] is the boundary between the loose powder and the solid part (see Figure 28, which shows the microstructure of a part cross section). Under certain conditions unconsolidated particles may remain in the solid part. This can be overcome easily by proper energy intake settings, correct scan strategies like cross scanning (in X, Y, in X&Y and so on) or multi times scanning of the same layers if required. The boundary conditions may be partially improved by the so called out-line scanning imposing a sharper border line between loose powder and consolidated part. A powder with slow recrystallization rate, as depicted by non-overlapping or slightly overlapping endothermic and exothermic peaks during the heating and respective cooling phases of the DSC analysis (Figure 26), might result in nearly fully dense parts with minimal distortion, while highly overlapping peaks will yield bad results (Figure 27) [212].

Figure 26 : DSC plots (heating and cooling) of new reinforced PA 12 powder containing elongated rather

than spherical beads for better strength

Figure 27 : DSC plots (heating and cooling) of IP60 powder showing overlap between melting and recrystallization peaks [US patent 5,648,450]

In contrast to metals, one has to be aware about the thermal deterioration of the powder material. The recyled material has a limited usability as demonstrated for PA 12. The long exposure to heat leads to chain growth, a rise in

-742-

MW and so a rise in viscosity. This causes nonconstant consolidation conditions and a shift in the melting temperature. The empirical melting temperature ramps up as a function of build height and age of the powder. In the extreme this can lead to creation of patterns on the surface known as « Orange Peel » because of its texture (see Figure 10b). This can be improved by mixing with virgin material (about 70% virgin – 30% reused).

Melt Flow Index (MFI) is flow in grammes that occurs in 10 minutes through a standard die of 2.095 ± 0.005 mm diameter and 8.000 ± 0.025mm in length when a fixed pressure is applied to the melt via a piston and a load of total mass of 2.16 kg at a temperature of 190°C (some polymers are measured at a higher temperature, some use different weights and some even different orifice sizes). MFI is an assessment of average molecular mass and is an inverse measure of the melt viscosity; in other words, the higher a MFI, the more polymer flows under given test conditions. Hence the MFI of a polymer is vital to anticipating and controlling its processing. Generally, higher MFI polymers are used in injection moulding, and lower MFI polymers are used with blow moulding or extrusion processes.

When polymer powder is reused in SLS, the MFI is already about three times lower than the virgin powder (Figure 29). This leads to higher viscosity, and sometimes the quality of the product becomes unacceptable.

Figure 28 : Boundary between loose powder and solid part

-20

0

20

40

60

80

1 6 11 16 21 26 31 36 41 46 51 56 61 66

Build (run)

MFI

QAMFILinear (MFI)

Virgin

Interior quality

Recycled

Figure 29 : Melt flow index of powder as a function of number of builds in which powder is used (Source

FHSG St. Gallen)

Figure 30 : Tensile strength and elongation of PA-GF polymers as function of number of build in which non-

refreshed powder is used

With PA 12, process stability and repeatability in the consolidation process can be obtained by keeping the MFI within a constant range and never working with virgin powder [119]. This phenomenon is amplified for the glass filled PA 12 material, leading to a rapid decline in mechanical properties, a clear evidence of different consolidation conditions (Figure 30).The aging and the resulting change in consolidation phenomena and final component properties are very minor with amorphous materials like PS and elastomeric polyester based material. To conclude on polymers, Table 2 gives an overview of some commercial available SLS polymers with their most important properties concerning the SLS process.

Material

Tg (Glass transition

temperature)(°C)

Tm (Melt temperature)

(°C)

MW(molecular

weight) (g/mol)

PA 6 50 230

PA 12 (41) 184 9.600

PCL -65 … -60 58 … 60 40.000 … 80.000

PEEK 143 340

Table 2 : Overview of commercial available polymers with most important properties for SLS

3.2 MetalsToday metals are the second most used material type in SLS/SLM. Almost all consolidation mechanisms listed in Figure 2 can be applied to consolidate metallic parts [164].

3.2.1 Solid state sintering (SSS) Theoretically SSS can be applied for consolidating metal parts by SLS [204]. Its application is however limited for two reasons: The process is difficult to control: the process transits

between no consolidation (at lower energy levels) and partial/full melting (at higher energy levels) within a very narrow process window (range), making it difficult to adjust the process parameters for real SSS.

SSS requires a long interaction time between laser beam and powder particles. This calls for slow scanning velocities and makes the process not

Un-moltenparticle core

Un-moltencompleteparticlestuck to edge

Fullymoltenparticle(no core)

-743-

economically viable. Tests performed at K.U.Leuven with steel powders indicated that scanning speed in solid state SLS needs to be limited [204], where those scanning speed may reach about 1500 mm/s in SLM and about 10000 mm/s in liquid phase SLS with Laserform [103].

Early SSS laser sintering tests with steel were performed by Van der Schueren [95, 204]. Gusarov has applied and modelled laser SSS of Ti powder [66, 67]. The laser interaction time was of the order of magnitude of 5 seconds, as compared to 0.1- 0.3 ms with other binding mechanisms. Tolochko sintered Ti teeth with a static defocused Nd:YAG laser beam [194]. The Ti parts obtained were dense in the core, i.e. in the laser spot (high energy density yielding SLM), but had a porous shell (lower energy yielding SSS). No post-infiltration was aplied to remove this porosity, as it favoured osteo-integration of bone into the Ti tooth. Lanzetta has experimented with gold powders [111]. He succeeded to thermally sinter pure Au spherical powder ranging between 5 and 40 μm in size and then to infiltrate the obtained skeleton with binary gold eutectics based on silicon, germanium and tin, with a melting point as low as 278° C. The initial thermal sintering is not yet laser-based, but this is the logical next step. This laser sintering could be done by solid state sintering or by liquid phase sintering (partial melting) as described further on (see §3.2.2, subsection B1).

3.2.2 Liquid phase sintering and partial melting This category unites many different kinds of technologies. Most of these techniques combine a structural material remaining solid throughout the process and a binder material being liquefied. In some cases, however, the solid and the liquid phases result from the same material. The

first group of technologies is characterized by a clear distinction between the binder and structural materials. A. Different binder and structural materials Technologies in this category can be further divided into three groups according to the type of powder grains that are used (Figure 2).A1. Separate particles These technologies use different binder and structural particles [4]. The structural material (in this case, a metal, alternatively a ceramic as discussed below) should generally have a higher melting point than the (metallic) binder material (Use of polymer binders has been discussed in §3.1). The binder particles are usually smaller than the structural ones, in order to facilitate their preferential melting. However, preferential binder melting may be counteracted by the higher reflectivity or lower laser absorption of the metallic binder material (typically Cu or Co) as compared to the structural material (metal or ceramic). In some cases, this might even lead to a reverse melting process, in which the structural particles will melt prior to the binder grains [105, 204].The combination of small binder particles and larger structural particles has the additional benefit of better packing with small pores, favoring fast spreading of the molten binder by capillary forces and fast rearrangement of the particles. Generally a ‘green’ part is produced which is still porous and brittle. Therefore, a post treatment consisting of a furnace post-sintering, Hot Isostatic Pressing (HIP) or an infiltration with a low melting point material (metal, alternatively epoxy or other) is usually necessary: see §4. To obtain sufficient mechanical properties, most commonly infiltration is applied because of its effectiveness and relative ease.

a) WC and Co powder mixture b) mech. alloyed WC-Co powder

c) sintered powder mixture d) sintered mech. alloyed WC-Co powder

Figure 33 : WC-Co powder mixture and sintered parts (source: University of Leuven)

Many different material combinations have been tested in the past. At the University of Leuven, metal-metal composites were tested as well as metal-ceramic composites [64, 100, 101, 109, 112]. Some examples of metal-metal composites are Fe-Cu and Stainless Steel-Cu. As for the metal-ceramics combinations (see also §3.3), WC-Cu, WC-Co, WC-CuFeCo, TiC-Ni/Co/Mo, ZrB2-Cu and TiB2-Ni were tested. Figure 31 and Figure 32show cross sections of Stainless Steel-Cu and WC-Co green parts (see also Figure 33 a and c). A2. Composite particles Composite powder particles contain both the binder and the structural material within each individual powder grain. The powder may be obtained by mechanically alloying a mixture of two different powders, causing powder particles to be repeatedly milled, fractured and welded together.

Figure 31 : LPS of Stainless Steel-Cu powder mixture (a: nonmolten steel particle, b: molten Cu, c: porosity)]

100 m

Figure 32 : LPS of WC-Co powder mixture; top: before infiltration (a: nonmolten WC particle, b: molten Co, c:

porosity); bottom: after infiltration with copper [112]

-744-

Figure 33 a shows an initial powder mixture of WC and Co particles. Mechanical alloying results in WC-Co powder particles with a micro grain composite structure in which the two different phases (WC and Co) can still be identified (see Figure 33 b). Such composite particles yield a higher SLS green density and a better surface roughness than a mixture of separate WC and Co powders (compare Figure 33 c and d) [112]. Figure 34shows an example of a WC-Co injection-moulding insert.

Figure 34 : Bronze infiltrated WC-Co injection moulding insert (source: University of Leuven)

A3. Coated grains A third possibility to combine a binder and a structural material is to coat the structural material with the binder phase. This ensures that the laser radiation hitting the powder particles is preferentially absorbed by the binder material that is to be melted. Moreover a more effective bonding of the structural particles is realized since the binder material already surrounds all structural particles. Coated powders exist both with metal and polymer binder coatings.The University of Leuven experimented with Cu coated steel powder which turned out to be successful despite the fact that the powder coating process was hard to control [204].Figure 35 shows a cross section of an SLS part made from commercially available polymer coated stainless steel powder (LaserForm), after polymer debinding and infiltration with bronze.

Figure 35 : Bronze infiltrated Laserform ST 100 part

At the University of Queensland, an aluminium LPS process has been developed in collaboration with 3D Systems [171, 172, 173]. Since the strong oxide layer on the grains prohibits direct sintering of Al powder, an indirect procedure was developed. Nylon coated Al particles are fused together using the LPS mechanism. Next, the nylon binder is de-binded in a furnace with N2atmosphere. A small amount of magnesium added to the powder mixture makes it possible to form aluminium nitride instead of aluminium oxide when heating up to 540 ºC. The aluminium nitride acts as a rigid skeleton, which is necessary during the infiltration phase. The infiltrant being used is a ternary eutectic aluminium alloy (Al-13.8Si-4.7Mg), having a lower melting point than the basic aluminum alloy. Relative part densities up to 95 % were obtained. Figure 36 shows the different processing steps. B. No distinct binder and structural material Technologies in this group do not exhibit a clear distinction between binder and structural phases (even in cases where different powders are mixed). Rather than the distinction between binder and structural material, there is a distinction between molten and nonmolten

material areas. Therefore ‘Partial Melting’ is a better name for these technologies than ‘Liquid Phase Sintering’. B1. Single phase, partially molten When the heat supplied is insufficient to completely melt powder particles, only a shell or small particles will melt. This way the molten material acts as a binder between the nonmolten particle cores [184, 195].The partial melting phenomenon was modeled by Karapatis [51, 52, 53, 82]. Using a simple thermal model, skin and core temperatures of the powder particles were calculated. This way the minimal pulse energy to fully melt the particle can be calculated. Below this value, the core temperature never exceeds the melting temperature and only partial melting is obtained. B2. Fusing a powder mixture Powders consisting of multiple kinds of powder particles can be classified in this group when they are only partially molten [64, 220]. For example at the University of Leuven experiments were done with a Fe-Fe3P-Ni-Cu powder mixture, aiming at the production of full density parts [93].The addition of a melting point lowering additive like Fe3Pis favorable in making the process more energy efficient. For instance, alloying pure Fe with a small amount of P lowers the melting point of pure Fe (1538°C) to the eutectic temperature of the Fe-P system (1048°C, see Figure 38). Moreover, the dissolution of P in Fe also has the benefit of a lower surface tension of the melt, resulting in a better wetting behavior. Ni is added for its marked strengthening effect. A micrograph of the structure obtained after partial melting of this powder is given in Figure 37. A closer look at the microstructure reveals that not all the powder particles are molten. The final part consists of a low melting point P-rich phase (no. 3), a high melting point phase with no significant amount of P (no. 2), some remaining porosities (no. 4) and some remaining unmolten Fe powder particles (no. 1). Therefore, the process cannot be called ‘full melting’ and the name ‘partial melting’ is preferred. Notice that, even though no full melting occurs, almost full density is achieved.

50 m 50 m

20 m

a b

cFigure 36 : Steps in Aluminium SLS: a) green part

(nylon binder in black, not visible) b) aluminium nitride skeleton surrounding the aluminium grains, c) infiltrated

part

Figure 37 : Micrograph of multiphase steel powder: 1) unmolten Fe particle, 2) high melting P-poor phase, 3)

low melting P-rich phase, 4) pores [93]

-745-

Figure 38 : Phase diagrams of Fe-P and Cu-P showing the temperature lowering effect of the added P

EOS’ Direct Metal Laser Sintering process (DMLS) also belongs to this category [45]. Figure 39 and Figure 40show the difference in porosity obtained with 50 μm and 20 μm DMLS powders. In both cases, shot peening enhances surface quality and the density of the top layer. Different parameter sets are used for the part’s skin and core in order to obtain a rigid and hard shell and at the same time a higher building speed (by fast scanning of the core).

3.2.3 Full melting Since a couple of years, there is a big move in the processing of metal powders from laser sintering or partial melting towards full melting (SLM) [25, 27, 94, 97, 99,146, 147, 156, 204] , The major benefits and drawbacks of SLM are given in Table 3.The binding mechanism invoked in full melting is largely driven by the fluid behaviour of the melt which is related to: Surface tension (a/o Raleigh instabilities) Viscosity Wetting Thermocapillary effects (a/o Marangoni convection) Evaporation Oxidation

Changes of viscosity across the melt pool, due to changes of viscosity between the liquidus and solidus temperature, may largely influence the shape of the tracks and the resulting smoothness and density of the parts obtained [158, 163].K.U.Leuven did an extensive study on the effect of alloying elements on binding and melt pool stability in SLM of ferro materials [158, 163]. Some major conclusions

are: Increased oxygen content leads to an enlarged melt

pool size due to exothermal oxidation of Fe. This deteriorates the surface quality.

Adding carbon (steel powder 99.1% Fe, 0.1% O, 0.8% C versus 99.9% Fe, 0.1% O) induces spherical pores, caused by entrapment of CO or CO2 gas bubbles. A reduction of the oxygen content in the SLM part is observed, together with a depletion of carbon. Carbon ameliorates the surface roughness of single layers, while worsening it on 3D parts.

200 m

100 m

Figure 39 : DMLS Direct-Metal 50V-2 grain size 50 μm, porosity 10-15%: a and c before shot-penning, b and d after shot peening (a and b: cross section, c and d: top

view) (Source EOS [46])

a b

c

1 mm

Figure 40 : DMLS Direct-Steel 20V-2 grain size 20 μm, porosity ~ 5%, in the ZX plane: left before and right after

shot-penning, cross section (Source IPT)

De-oxidizers like Si and Ti do not improve the process. On the contrary, they lead to a larger melt pool and higher tendency to balling, resulting in bad surface quality. They reduce the amount of spherical pores, but increase the amount of irregular pores. Cu does not have a large effect on the melt pool behavior, however when added in pure form, it reduces the absorption of laser powder due to its high reflectivity.

The University of Liverpool studied the effect of small additions of boron, in the form of iron boride, to steel [23].

Full Melting Benefits Drawbacks Material choice No distinct binder and melt phases; hence,

the process can produce “single material” parts (e.g. Steel, Ti or Al alloys), rather than producing a composite green parts which might not be desired

Not suited for well controlled composite materials (e.g. WC-Co)

Production steps (time, cost)

Elimination of time consuming and costly furnace post-processes for debinding (in case of polymer binder phase), infiltration or post-sintering

The laser powder processing needs higher energy level: i.e. high laser power, good beam quality (more expensive laser) and smaller scan velocities (longer build times)

Part quality Better suited to produce full dense parts (even over 99.9%) in a direct way, without post infiltration, sintering or HIPing

SLM suffers more from melt pool instabilities (low quality of down facing surfaces, higher upper surface roughness, risk of internal pores) and higher residual stresses (common need to build and anchor part on solid base plate, risk of delamination, distortion when removing base plate).

Table 3 : Benefits and drawbacks of SLM

d

-746-

The boron is aimed to replace interstitial carbon in steel, which does have a negative effect on SLM processibility. The addition of boron did not have such deleterious effect on processibility and to some extend increased the part density. Addition of Ti to the Fe-FeB powder confirmed the negative effect of Ti on processibility, even though it improved the bonding of the samples to the base plate and reduced the extreme hardness. Notice that the results reported are not always consistent and may differ in view of the machine or laser used, the layer or part height, etc. Decades of research might still be needed to develop appropriate metallurgical understanding and tuned ferro materials for SLM, in a similar way as done in the past for hot or cold working steels, free-machining steels, high speed steels, etc. It is not surprising that development and commercial release of ferro powders for SLM happens at a very low rate, and that their number to date is so limited (mainly Fe-based powders and stainless steel; very few tool steels), not withstanding the large research effort deployed [35, 73,145].The move from SLS to SLM represents a major advance in rapid manufacturing of non-ferrous metal parts, which to date could not be well processed using SLS solid state sintering, liquid phase sintering or sacrificial polymer binders [109, 138]. SLM of powders from Ti or CoCr alloys has entered industrial practice, mainly in the field of medical applications: e.g. for dental crowns, copings and frameworks [2, 15, 85, 104, 113, 188]. Those powders can be melted more easily to almost full density than ferro powders (Figure 41). This figure demonstrates that the density of Ti6Al4V is highly repeatable and controllable through the processing parameters up to densities of 99.98% [205]. SLM of Al might be the next material reaching industrial application [43, 173]. The University of Liverpool tested SLM of Cu [149], while the work of Lanzetta [111] might be a first step towards SLM of gold.

3.2.4 Chemically induced consolidation Metals and hardmetals (cermets) can also be consolidated chemically. In situ Cu-based composites can be synthesized by reaction between elemental Ti and C powders in Cu. Using a CO2 laser, Cu can be melted with the help of Ti and the reaction heat of the TiC formation [125].The example given above shows a possible way to apply chemical bonding in SLS of Al powder. In this case, reaction of Al with the N2 atmosphere of the SLS chamber can be used to create an AlN binder phase that binds the Al particles together (Figure 36b). To date however, this chemical induced binding mechanism has only been invoked during furnace post sintering [176].

3.3 Cermets and hardmetals Cermets and hardmetals, being composites in which ceramic particles are embedded in a metal matrix (binder), can be readily processed by liquid phase SLS using a mixture of ceramics particles that remain solid throughout the process and metal particles that are melted by the laser [36, 39, 206, 210]. In most of the cases, the sintered part will not reach full density and a post-infiltration of the porous green part will be needed (see Figure 32) [132].Many researchers investigated liquid phase SLS of common WC-Co cemented carbide (hardmetals) using either a mixture of WC and Co powder particles or composite WC-Co powder particles [112, 132]. Some examples and details, originating from K.U.Leuven, have been given under §3.2 “Metals”, subsection A (see Figure 32 to Figure 34). Researchers also investigated cermets like WC-Cu, WC-CuFeCo, TiC-Ni/Co/Mo, ZrB2-Cu and TiB2-Ni. More examples are given in § 3.4 ”ceramics”.

The Fraunhofer Institute of Production Technology (IPT) worked with different compacted, respective ground, WC-Co powder particles: see Figure 42. A higher mean density could be achieved when applying the spherical particles of the compacted WC-Co powder, in comparison with the irregular particles of the ground WC-Co powder. The metallographical analysis of laser sintered WC-Co structures showed a structure which is comparable with conventionally sintered hard metal. The distribution of the WC particles is even and dispersed within the cobalt matrix (Figure 43). This effect could be investigated because of a rearranging of the WC particles within the fluid cobalt phase. The viscosity of the liquid cobalt phase and the wetting behavior of the WC particles with the cobalt binder are sufficient for forming an almost dense structure. Inhomogeneities, caused by the layered structure, were not identified. The maximally achieved relative SLS (green) density when applying WC-Co 75-25 powder was 78.3% and 68% in case of WC-Co 88-12. The energy input per unit area is particularly important in terms of localized material densification (localized liquid phase sintering). In case of a constant energy input per unit area, a higher relative density was achieved by a fast scan velocity (vS) and an increased laser power (PL) in comparison with a slow scan velocity and a low laser power. Identical energy input per unit area resulted in an 10% increase of the relative density by the selection of a high vS-PL combination. Figure 44 shows parts made of WC-Co 88-12. Other researchers applied a sacrificial polymer binder(MMA-BMA) to consolidate ceramic particles (e.g. SiC) into a green part [11, p. 111]. This green part has then to be polymer debinded in a furnace, after which a further furnace consolidation might be applied (mostly infiltration with copper or bronze). An example of production of cermets by chemical induced consolidation is reported in §3.4.7, where Al2O3cermet is produced by making CuO react with Al. Free Cu is added to control the exothermal reaction [80].

Figure 41 : Density of SLM processed titanium with different scan spacing and scan speed

3.4 CeramicsBesides polymers and metals, ceramics is another important group of materials with many applications in the field of mechanical engineering. Therefore the qualification of ceramics for rapid manufacturing technologies is a prior objective of worldwide research activities. This is an ambitious challenge, because the specific properties of most ceramics conflict with

-747-

technological material requirements for SLS/SLM (e.g. low thermal conductivity, high melting point etc.). By now, technologies of laser and powder bed based layer manufacturing using ceramics are not so well-investigated as compared to the technologies using polymers or metals. However there are some research activities which differ in various technological aspects. A differentiation can be made by the type of material, the type of consolidation and the necessity of post-processing of the considered technology.

3.4.1 Types of Technical Ceramics Ceramic materials are inorganic and non-metallic. Technical ceramics means ceramic products made for technical applications. Usually they are shaped from the "green body" at room temperature and acquire their typical properties during a long sintering process at high temperatures. In laser and powder bed based layered manufacturing of ceramics the very short time of heat impact is the most critical point. The different sintering mechanisms (described in §2) are more or less time-dependent. In accordance with their chemical composition the technical ceramic materials can be divided into three main groups:

silicate ceramics oxide ceramics non-oxide ceramics

Silicate ceramics are multi-phase materials; the fundamental ingredients are clay, kaolin, as well as feldspar and soapstone as silicate carrier. Further ingredients such as alumina and zirconium silicate (ZrSiO4) are added to achieve specific properties. These materials, obtained from natural raw materials, combine the basic electrical, mechanical and thermal properties of technical ceramics. The predominating types of powder consolidation mechanisms are liquid phase sintering, partial melting and full melting. Amongst the materials of silicate ceramics are:

technical porcelain steatite cordierite mullite-ceramic

Oxide ceramics consist of at least 90% of single phase and single component metal oxides. These materials are glass-phase low or glass-phase free. Synthetic raw materials with a high level of purity lead to an even structure with very good properties at a very high sintering temperature.Various raw materials from mining areas in the whole world form the basis for the production of high-quality oxides and mixed oxides. The predominating types of powder consolidation mechanism are solid state sintering, partial melting and full melting. Following materials belong to the group of the oxide ceramics materials:

aluminium oxide magnesium oxide zirconium oxide aluminium titanate piezo ceramic

Non-oxide ceramics are material compounds of silicon and aluminium with nitrogen or carbon. In general non-oxide ceramics demonstrate a high share of covalency bonding which provides them very good mechanical properties, even when being used at high temperatures. The predominating types of powder consolidation

mechanisms are chemical induced binding, partial melting and full melting. Amongst non-oxide ceramics are:

Carbide Nitride

Figure 42 : WC-Co powder particle

Figure 43 : Distribution of WC particles in Cobalt matrix

Figure 44 : Parts made of WC-Co 88-12.

Selective laser consolidation of ceramic powders has already been investigated for various materials and various applications: e.g. micro SLS of SiC powder [154,187], sintering of transparent Ta2O5 dielectric ceramics [78, 79], production of SiO2 investment casting shells [88],production of biocompatible medical implants [50, 115,160, 202, 213], SLS of bismuth titanate ceramics [131],SLS of radiation detectors using bismuth titanate (Bi4Ti3O12) ferroelectric ceramics [130] and bismuth germanate ceramics (Bi4Ge3O12) [129] etc. Other examples are given below.

3.4.2 SLS of ceramics using polymer binder One way to powder consolidation is the use of a sacrificial polymer binder [8, 71, 72, 140]. The used powder material consists of a ceramic and polymer powder mixture or of polymer coated ceramic particles. The polymer fraction of the material melts within a short time and at low temperatures and binds the ceramic particles in a polymer matrix. The mechanical properties of such ceramic parts are not sufficient, so that in most cases a post-consolidation process combined with a debinding, e.g. furnace treatment, is necessary.

-748-

The University of Liverpool has developed the SLS of SiC/Polyamide Composites [76]. In this process, the powder is a mixture of SiC and PA particles. Unlike the majority of the previous research in this area, the polymer remains an integral part of the final product rather than being removed in downstream processes. The evaluation of the mechanical properties of the sintered composite parts is therefore an integral part of the research. The materials used were FEPA (Federation of the European Producers of Abrasives) standard silicon carbide grits individually blended with Duraform Polyamide which is based on nylon 12. SEM micrographs of the fracture surface of SLS parts are shown in Figure 45.

Figure 45 : SEM micrographs of the fracture surface of SLS parts build at 8 Watts laser power [76]

Other polymer coated ceramic powders are the sand powders developed by DTM-3D Systems (SandForm) as well as by EOS GmbH (Direct Croning Process) and that have found industrial application for some decades now [21, 183, 193]. The powders offered are standard Si or Zr sands, coated with a small phenolic layer, responsible for fusing the sand grains together. These powders are used to produce moulds and cores for metal casting. Stabilized zirconia powder in mixture with a 80:20 PMMA-BMA copolymer binder was used at the University of Texas in Austin to produce shells for titanium casting by SLS, because of its low reactivity and its thermal shock resistance. A human femur head was reproduced starting from laser-scanned data, out of Ti-6Al-4V to demonstrate the possibility of the technology [71].

3.4.3 Liquid phase SLS of ceramics If multi-phase ceramics are used for SLS, the dominating powder consolidation mechanism is the liquid phase sintering. It always occurs if one phase of the material is already molten while the others are still solid. A great advantage of the liquid phase sintering is that the molten phase can bind the solid phase in a matrix structure within a short time [117, 124]. Unlike solid state sintering where the diffusion process takes a long time, the SLS process can run faster. The Fraunhofer IPT began its investigations into ceramic laser sintering in 1997 using laser sinter equipment with maximum 100 W laser power (CO2 laser) [88]. Initial tests performed on different ceramics such as zirconium silicate or aluminium silicate showed that the zirconium silicate was more suitable. Fundamental experiments were performed using ground powders of differing fineness and mixtures of standard grindings. The results showed that higher percentages of dust led to a superior surface quality and to higher sinter rates while all other parameter combinations remained constant. This powder was not, however, suitable for use in an automated process, as it could not be layered evenly. This, in turn, was due to the powder heavily agglomerating and sticking to the wiper. The problem was solved by fractioning off the dust from the powder. Alternatively, one can use spray granulate: the almost spherical particles make automated spreading possible, even with small sizes. Grains of less than 50 μm are currently in use. The effect of each process parameter on laser sintering is determined by varying each one

separately. An increase in laser power makes the sample denser. The introduction of more energy into the powder leads to an increased liquid phase, which seals off the porous structures of the layer below. The surface is, however, significantly rougher due to the melted particles joining together into large drops (balling see § 2.3). This can lead to the wiper ripping off a previously sintered layer as it tries to wipe on the next layer. An increase in scan velocity leads to the opposite effect: a shorter exposure time means that less energy is introduced into the powder. This reduces the amount of material in the liquid phase, thus lowering component density but improving surface quality. Small scan-line spacings (high overlap and energy) also result in balling, while scan spacings greater than the focal diameter result in very fragile parts. The maximum density of the solid and manageable components produced in a stable laser sintering process was approximately 50 % of the theoretical density. The results of further tests with silicon dioxide (SiO2) also proved positive and allowed to produce investment casting cores with good strength, allowing the liquid wax to be injected around the laser sintered cores as is common in conventional core production [88].

3.4.4 SLS of ceramics slurries The powder based process can be substituted by a ceramic slurry based process to improve the green density of laser sintering parts. The ceramic slurry allows the use of smaller ceramic particles which leads to a higher green density in every single layer. At Fraunhofer IPT, the SLS of ceramic slurries was developed by Klocke [86]. Compared to former investigations on ceramic laser sintering, where manufacturing of powder-based sintering was investigated, the aim of his work is to provide a knowledge base for layer manufacturing of ceramic suspensions by laser energy. By the use of ceramic slurry a higher green body density in the dried stage is available, compared to powder-based laser sintering. Thus, a higher density of the final part is suggested. Figure 47 shows the process steps in suspension laser sintering. Besides the three process steps of (conventional) powder-based laser sintering, »coating«, »exposing« and »platform lowering«, the suspension laser sintering process has one additional step, »drying«. The re-coating is equivalent to the spraying of a ceramic suspension layer. Before exposing the layer, the slurry must be dried completely. The laser sintering of ceramic suspensions is subject to several effects which define the quality of laser sintered parts. Depending on the cause of appearance they can be clustered to material-, process- and machine-effects. The result is evaluated by both the part quality and the process stability. Modelling of Suspension Laser Sintering The following model of the suspension laser sintering process helps to describe and to illustrate the mathematical theories of single and overlapping process parts. Sun divided the conventional powder-based laser sintering into three model parts: the optical, the thermal and the sinter model [190, 191]. Below a new drying model, which is essential for the suspension laser sintering, is described. The drying model is based on the results of an analysis of the system. Subsequently the physical phenomena are described regarding the boundary conditions and expressed by mathematical equations. The wetting behaviour is an important part of the drying model. Solid parts tend to accumulate other matters. The accumulation of gaseous molecules to solid or liquid surfaces is called adsorption. The system solid-liquid is called wetting. The

-749-

dimension of the wetting depends on the combination of materials [161]: see also § 2.3. Different cases of wetting which are possible for suspension laser sintering are illustrated in Figure 46.The contact angle describes the wetting behavior. If the contact angle > 90°, the system is called not wetting. A wetting system has a contact angle < 90°. The best wetting behavior, which is desired for suspension laser sintering, exists for = 0°. xx is the surface energy or the boundary energy of the different materials. The following equation applies to a state of minimized energy [161].

lv

slsvcos (2)

A dried layer consists of single particles. At the interface of two particles exists an other boundary energy ss. This leads to a three-phase interface at the end of the solid-solid interface. Figure 46 d and e show this case. The following equation represents the boundary energy of the particles.

2cos2 slss (3)

Figure 47 : Process steps of suspension laser sintering

The intended wetting behavior is set up by variation of sl.The use of adequate suspension additives results in a reduced value of sl, which is necessary since the value of

ss is low for similar particles in the dried layer. This leads to an acute angle of . Another important element of the drying model is the drying or evaporation process. The drying process consists of thermal separation of a liquid phase out of a powder material [57].A green part can only be laser sintered, if the residual water content is less than 3 wt-% [57]. Otherwise the part could be damaged by abrupt vaporization of the liquid phase during laser exposure. In the suspension laser sintering process air is used as drying agent. This means that the liquid phase of the suspension must exhaust to ambient air during the drying process. This is only possible with unsaturated air. Therefore the humidity L as

well as the moisture content x are important effects of the drying process. They were calculated by the following equations.

DS

DL p

p (4)

L

Wmm

x (5)

In these equations pD is the partial pressure of the water vapor, pDS is the saturation vapor pressure, mw is the mass of water and mL is the mass of dry air. The humidity absorption of the air increases with rising temperature. If the water content x is smaller than the saturation water content xS, the desired drying can proceed. This effect can be amplified by a continuous supply of dry air. Another important part of the drying model is the self adjustment of particles during the drying process. An ideal packing density can be accepted [92]. Figure 48 shows the self orientation of particles by capillary forces. F is the lateral capillary force and Fd the hydrodynamic force which drives the particles in the direction of thinner areas. Fd is caused by the hydrodynamic flow Jw which compensates the vaporization of the liquid. The accepted ideal packing density supports the thesis that the use of a ceramic suspension instead of a ceramic powder reaches to a higher density of the green body before laser sintering.

JW

FF

Vaporization

FdFJW

FF

Vaporization

FdF

Figure 48 : Self orientation of particles by capillary force Laser Sintering of Zirconia (ZrO2)Investigations were made with a selection of different oxide ceramics suspensions. In this paper the results of the laser sintering of zirconia are discussed as an example. It is a suspension of yttria–stabilized zirconia and water. The slurry is made of spray granulates (TZ-3YS-E) by Tosoh Corporation, Tokyo, Japan. This material has an addition of 0.25 % alumina. This leads to a sintering of the material at a lower temperature. Figure 49 shows the sintering temperatures of two different zirconia materials [198]. The E-Type needs a sinter temperature that is about 100 °C lower than the pure material. 3YS-E starts sintering at 1350 °C and 3YS at 1450 °C. The parameters used in this investigation are shown in the table in Figure 49. One way of sintering zirconia is to use a low scan velocity combined with a low laser power. The influence of laser power and scan velocity on the surface roughness can be seen in Figure 50. The roughness decreases with decreasing laser power and increasing scan velocity. However the strength of the parts decreases in the same way: the parts in the lower right side were not built-up completely. A high roughness leads to contacts with the wiping system during the process, so that the parts in the upper left were not built-up

Figure 46 : Wetting behavior of the solid-liquid system and interfacial equilibriums of a solid body with two grains

-750-

completely, too. The other three parts which lay on the diagonal from the lower left to the upper right corner were built-up completely in 25 layers. The lower pictures in Figure 50show the green (left) and solidified (right + enlargements at the bottom) surface of the part from the upper right corner (11W, 200 mm/s). At the right magnification the grain boundaries are visible. The polycrystalline structure could be generated by a solid phase sintering process. Cracks can be seen in the picture in the lower right corner. They are caused by contraction during the cooling phase. Figure 51 shows the density and the open porosity of samples from a test series. The tests are made with a layer thickness of s = 0.1 mm. The hatch space is hS = 0.01 mm and the laser power is PL = 13.0 W. The scan velocity varied from 50 mm/s up to 150 mm/s. Density generally decreases with increasing the scan velocity. No trend in open porosity is visible. This can be explained by a changing ratio of open and closed porosity.

6.0

5.9

5.8

5.7

5.6

5.5

1200 1300 1400 1500 °C

TZ-3Y-E

TZ-3YS-E

TZ-3Y

TZ-3YS

, Density6.1 g/cm3

6.0

5.9

5.8

5.7

5.6

5.5

1200 1300 1400 1500

6.0

5.9

5.8

5.7

5.6

5.5

TZ-3Y-E

TZ-3YS-E

TZ-3Y

TZ-3YS

1200 1300 1400 1500 °C

, Density6.1 g/cm3

6.0

5.9

5.8

5.7

5.6

5.5

1200 1300 1400 1500

Figure 49 : Sintering temperatures of different zirconias [198] and an overview of process parameters

Other investigations in the field of SLS of ceramic slurries were made at TU Clausthal, Germany [32, 33]. They developed the layer-wise slurry deposition (LSD) which can be used to fabricate hydroxyapatite (HAP) bone implants. HAP suspensions with a maximum solids loading of 66 wt% were prepared, and the compositional and structural changes during laser treatment were studied. The optimized slurries have a slight shear thickening behavior and a low viscosity at pH 11.5. The rheological properties of the HAP slurries play an important role in facilitating the layering operation. The most important process parameters affecting microstructure and composition were laser power and laser beam speed. The penetration depth of the laser increased with increasing laser power and decreasing beam speed. In the LSD process, the largest layer thickness is limited by the penetration depth. When the laser beam speed was varied in the range of 10-60 mm/s and the laser power in the range of 18-24 W respectively, the maximum layer thickness was 100 micron. The energy

density was estimated between 400 and 1800 J/cm2. The sintered samples had a pore size of 40-70 μm. After laser sintering, the presence of alpha-tricalcium phosphate and amorphous calcium phosphate was detected [160].

11.0 W

200 mm/s

6.8 W

10 mm

50 mm/s 100 mm/s

9.0 W

11.0 W

200 mm/s

6.8 W

10 mm

50 mm/s 100 mm/s

9.0 W

2 μm2 μm2 μm2 μm2 μm2 μm2 μm2 μ

5 μm 20 μm

200 μm200 μ200 μ200 μ2 μm2 μm2 μm2 μm2 μm2 μm2 μm2 μ

5 μm 20 μm

200 μm200 μ200 μ200 μ

Figure 50 : Green and laser sintered surface of TZ-3YS-E (PL = 11.0 W, vS = 200 mm/s)

vs = 50 mm/s

vs = 120 mm/svs = 150 mm/s

vs = 80 mm/svs = 50 mm/s

vs = 120 mm/svs = 150 mm/s

4.3

4.4

4.5

4.6

4.7

4.8 g/cm 3

20.0

22.0

24.0

26.0

28.0

30.0

32.0 %. Open porosity POpen

20.0

22.0

24.0

26.0

28.0

30.0

20.0

22.0

24.0

26.0

28.0

30.0

20.0

22.0

24.0

26.0

28.0

30.0vs = 50 mm/s

vs = 120 mm/svs = 150 mm/s

vs = 80 mm/svs = 50 mm/s

vs = 120 mm/svs = 150 mm/s

4.3

4.4

4.5

4.6

4.7

4.8 g/cm 3

20.0

22.0

24.0

26.0

28.0

30.0

32.0 %. Open porosity POpen

20.0

22.0

24.0

26.0

28.0

30.0

20.0

22.0

24.0

26.0

28.0

30.0

20.0

22.0

24.0

26.0

28.0

30.0

Figure 51 : Density and open porosity of TZ-3YS-E

Further, TU Clausthal uses the LSD process to present a SFF technique for ceramics in the alumina–silica system by layering binder-free, high-loaded ceramic slurries, followed by selective laser sintering. The low melting silica phase and the reaction sintering between silica and alumina favor the rapid prototyping of pure ceramic parts. On the basis of electroacoustic and viscosity measurements, stable slurries from Al2O3/SiO2 powder mixtures and water with a high fluidity have been prepared for the layer deposition with a doctor blade like in tape casting. Layers with a thickness of about 100 μm were processed. It was found in laser parameter studies that ceramic parts can only be obtained using special alumina contents and laser parameters. But the biphasic approach may allow greater flexibility in the processing regime than is afforded by the use of just one material. The microstructure of these parts depends mainly on the temperature gradient induced by the laser absorption and thermal conduction. The wet shaping facilitates laser-sintered parts with a relatively high density, which could be

-751-

increased by a thermal post-treatment [56]. The scheme of the LSD process is shown in Figure 52.

3.4.5 Solid state SLS of ceramics Phenix Systems (France) commercializes an SLS-like system that realizes SSS using a high temperature process chamber reaching up to 800 C [13]. The compacted powder, which is pre-heated close to the SSS temperature, is sintered due to the extra energy contribution of a Nd:YAG laser source. This way, ceramic materials can be processed as well as metal powders. To obtain the desired characteristics, a post sintering operation is necessary. Figure 53 shows some ceramic parts.

3.4.6 SLM of ceramics Fraunhofer ILT is currently developing a new approach in the field of SLM of ceramics [213]. It is based on direct and complete melting of a ceramic powder material by means of a laser beam. The powder material used consists of zirconia with small additions of other oxide ceramic materials. It contains neither any glass nor any metallic component.

Figure 52 : LSD main process [56]

Figure 53 : Some ceramic parts by Phenix Systems [13]The cross section in Figure 54 demonstrates that there is almost no porosity in the part. The high density results from the complete melting of the powder material and is achieved without any post-processing. However, all manufactured parts contain a large number of micro-cracks as shown in Figure 54b. The cracking occurs during the build process and is caused by thermal stresses.

The mechanical strength of specimens created by SLM was tested using the four-point bending method. Five specimens were tested and an average bending strength of 9.79 MPa was measured. This value is very low compared to a strength of more than 1000 MPa achieved with this material using conventional manufacturing techniques. The low strength is due to the large number of micro-cracks that are quite evenly distributed throughout the specimens. To improve the mechanical strength, current work is focusing on avoiding crack formation during the manufacturing process. It is expected that the cracking can be avoided by heating the build chamber to a sufficiently high temperature during the build process. For

this purpose a high temperature processing chamber will be designed and built up at Fraunhofer ILT.

a

bFigure 54 : Cross section of conventional manufactured

ceramic specimen showing very high density (a) and cross section through a SLM manufactured ceramic specimen showing cracks (b) [213] (images: optical

microscope).

Figure 55 : SLM part and its cross-section [50]

(b)

Furthermore Fraunhofer ILT uses its SLM process to process a β-TCP/glass composite powder material. It has been shown that the Selective Laser Melting process is suitable to manufacture three dimensional freeform parts from this class of material. The resulting parts have a good resolution and accuracy as well as a high content of the bioresorbable TCP phase. Therefore the process has a good potential for the manufacture of customized bone tissue scaffolds for clinically desired implants for regenerative therapy. Improvements of the mechanical properties may be achieved by modification of the composite powder material and by mechanical

-752-

compaction of each powder layer during the SLM process. Within the scope of a recently started joint project the pore microstructure of the SLM generated components will be optimized, especially by modifying the granule size distribution. Furthermore a bioactive glass for the TCP/glass composite with a clinically made-to-measure resorption capability will be developed and the SLM process will be adapted to the optimized material. Figure 55 shows a SLM-manufactured part and its cross-section [50]. Selective laser melting has also been applied for the production of ultra-thin nonporous silicon layers [42].3.4.7 Chemical-induced and Self-induced SLS of

ceramics The Fraunhofer Institute IPT (Aachen, Germany) has produced SiC ceramic parts using a kind of laser chemical-induced binding [88]. No binder elements were used and the laser-material interaction times were very short, thus excluding the diffusion processes occurring in Solid State Sintering. When heating the SiC particles to a very high temperature, partial disintegration of the SiC into Si and C occurs. The free Si forms SiO2, which acts as a binder between the SiC particles. The parts are thus composed of a mixture of SiC and SiO2. Afterwards an infiltration step using Si yields full dense parts. Next to SiC, also other ceramics were tested, e.g. Al2O3 and ZiSiO4. Figure 56 shows a ceramic casting shell produced by SLS. The Osaka Sangyo University tried to produce cermets (Al2O3-Cu), ceramics (MoSi2), and intermetallics (TiAl) by SLS induced chemical reaction [80]. By adding chemical reaction heat to laser heat, the particles of the products of the reaction were bonded together by relatively low laser energy. The Al2O3 for instance is obtained by chemical reaction of CuO with Al. The exothermal heat of the chemical reaction enlarges the laser heating, thereby allowing the use of lower laser power (typically 3-6W or 2W/mm/sec). In order to keep the self propagating reaction under control, inert substances were added to the raw powder materials: e.g. Cu was added to the mixture of CuO and Al, or TiAl was added to the mixture of Ti and Al. The researchers from Osaka also added glass powder to Ti and Al as an inert element aimed at controlling the self-propagating TiAl reaction while acting as a binder phase for the TiAl. Some investigators invoke self-propagating high-temperature synthesis (SHS) to produce ceramic parts. Slocombe e.g. has produced TiC-Al2O3 ceramic parts by mixing up TiO2, Al and C powders and laser inducing an exothermal reaction [182]. Further research is reported in [189].

Figure 56 : SLS-produced investment casting shell for impeller wheel (source: Fraunhofer Institute IPT)

3.5 Composites Selective laser sintering/melting is well suited to manufacture parts in composite material by applying powders that contains the two different materials to compose the composite. This can be used to produce polymer-metal, polymer-ceramic, metal-metal or metal-ceramic composites. Several examples have been given

in the previous sections. The present section will focus on a few applications that do not directly fit and were not discussed in the sections on polymers (Reinforced polymer were already described in §3.1.4), metals (Metal-metal composites were described in §3.2.2), cermets and hardmetals (described already in §3.3) and ceramics (described already in §3.4). A typical example of a composite that can be produced by SLS is PCL/HA, where hydroxyapatite (HA) has been added to polycaprolactone (PCL) for enhancing its strength and biocompatibility [215].In SLS/SLM, composites have been manufactured mainly by using three methods: taking materials or material mixtures as required in the

end product [90, 91, 206, 207], e.g. formation of WC-Co composite from a powder mixture of WC and Co [112, 132],

using in-situ reaction, e.g. formation of Cu-TiC-TiB composite from a powder mixture of Cu, Ti and B4C[118],

using furnace treatment, e.g. formation of Al/AlN composite by post-processing in an oven in the presence of nitrogen gas and infiltrant Al alloy [173].The various types are further elaborated in the coming subsections.

3.5.1 Manufacturing composites by using two-component powders

This is the most common method used for consolidating a composite. Most noticeable application is the formation of Polymer Matrix Composites (PMC). The binding mechanism involved is LPS. Polymer powders and ceramic are mixed and laser sintered to form the composite, e.g. PCL and HA [215], PEEK and HA [192],PE and HA [70, 165], PA and nano-clay [84], PA and HA [166, 167] etc. [29, 211]. It has been found that PMMA-coated HA powder could facilitate better consolidation of the composite, but uncoated powder is used for the reason of biocompatibility. Coating of the powder has certain other advantages. In case of a nano Al2O3 and PS composite, nano Al2O3 particles are coated with PS to prevent the agglomeration of nano particles resulting in an uniform distribution in the polymer matrix [219]. The reinforced powder is used in the form of a particulate, because the fibre as a reinforcement otherwise gives rise to problem in layer flow and is not helpful in increasing the final density and strength [11]. Instead of taking a mixture of polymer powder and the reinforcement powder, a single composite powder has also been used, e.g. glass-filled PA powder, Al-filled PA powder [1, 46, 133], HA-filled HDPE powder [165], HA-filled PA powder [166, 167], TCP-glass powder [213] etc. Composite particles, instead of powder mixtures, help to overcome the difficulty associated with mixing the powders and yield uniform spread of composite components in the final product. However, if one of the components of the composite powder is a fiber, problems of consolidation occur [11]. The method has also been used to make Metal Matrix Composite (MMC), e.g Fe and graphite [180], WC-Co [112, 132], WC-Co and Cu [63,65], Fe, Ni and TiC [55] etc. and Ceramic Matrix Composite (CMC) [50, 213].SLS of ceramic and metal, in general, does not give as dense products as is found in the case of PMC. In order to fully consolidate it, extra material is added to the powder mixture [65] or the porous product is infiltrated [112,132].Addition of La2O3 to the mixture of WC-Co and Cu has decreased the surface tension of the melt and consequently increased its sinterability [65]. In other cases, consolidation of WC and Co green product has been done by infiltrating it with bronze [112,132].

-753-

3.5.2 Manufacturing composites by using in-situ reactions

Laser-induced chemical reactions are used to create in-situ particles during laser sintering. The laser beam can be used in two ways: to overcome the activation energy of the reactants and

to form chemical compounds and to trigger a chemical reaction which will not only form

the compound but generate enough thermal energy to propagate chemical reactions.

Both types of in-situ formation of compounds are better than the pre-addition of compounds because of following reasons: fine and uniform distribution of compounds better wetting release of exothermic energy helping consolidation.

Examples of first type are the formation of Cu-based MMC reinforced with TiB2 and TiC from a powder mixture of Cu, Ti and B4C [118]. The second type is a case of Self-propagating High-temperature Synthesis (SHS), e.g. formation of TiC-Al2O3 composite from TiO2, Al and C [182]; Al2O3-Cu composite from CuO and Al [80]; NiTi-HA composite from Ni, Ti and HA [178,179]. In the latter case, it has been found that though the product is porous, it could be suitable as an implant material. Chemical reactions also helped creating a binder phase in the case of laser processing of SiC, where disintegration of SiC and subsequent reaction with O2 gas forms SiO2 which binds the remaining SiC powder particles [88]. This consolidation mechanism does not yield strong parts and has not been further used. However, the properties of the parts have been enhanced with the help of infiltration. This will be described in the following subsection.

3.5.3 Manufacturing composites by using furnace treatment

Post-processing of laser sintered materials in a furnace is another way for manufacturing a composite [116]. It has been mainly performed in two ways: by using the furnace for debinding and/or infiltration, by using the furnace for chemical reaction and

infiltration.Examples of the first type are treating the laser sintered acrylic/glass product in a furnace where the binders get debound and glass partially changes to ceramic making a glass-ceramic composite [34, 61]. In this case, composite was manufactured using a furnace without using infiltration. In another case, a laser sintered polymer-coated SiC powder was infiltrated with Al in a furnace to form MMC [8]. The second types are the most versatile with the most used example being the formation of Si/SiC composites. Laser processed SiC is treated with a phenolic resin. The resin after curing in a furnace gives rise to carbon which reacts with some of the infiltrant Si to form SiC and produces finally Si-SiC composite. The amount of SiC in the consolidated composite could be controlled by the degree of treatment of the green product with phenolic resin [47, 117]. Another noticeable example of chemical reaction inside a furnace is the use of nitrogen gas for reacting with the green product of Al alloy to form AlN [173].

4 POST-CONSOLIDATION Several of the selective powder consolidation techniques described above allow obtaining the desired final part (with full density or controlled porosity) directly, i.e. without need of post-consolidation. This typically holds for selective laser sintering of polymers and selective laser

melting of metals. However, in several cases, a furnace post-consolidation is needed, like post-sintering, infiltration [5, 44, 143] or HIPing [37]. Post-consolidation may be needed as well for intentionally porous final parts (to reinforce necks and part strength) or for intended fully dense final parts (increase density, strength, surface smoothness, etc.). Post-consolidation is not treated in detail in this paper: it partly consists of common PM technology, even though important new developments emerged specifically related to post-consolidation for layered manufacturing [17, 68, 186]. Important in the present context is to know that, if post-consolidation is to be applied, the primary selective laser consolidation process has to account for it. Some examples: If infiltration is to be used, the green part will have to

demonstrate (sufficient) open porosity. Experiments have proved that a higher green density will prohibit further densification by infiltration. A higher green porosity is required in this case and laser processing parameters will have to be chosen accordingly [106]. A major issue is also to find the proper match between laser processed powder and infiltration materials. A typical example is the work of Sercombe on SLS of aluminium: a first PA-coated Al powder (alloyed with Mg to avoid oxidation that would hamper infiltration) is used during SLS, while another eutectic Al powder is applied for infiltration [171,172]. Infiltration is not only used to consolidate SLS part; it has also been used to enhance e.g. their optical properties [143].

On the contrary, post-sintering will only be possible if the green part has a sufficient high density, which might be open or closed. In many cases the green (i.e. after SLS) or brown (i.e. after polymer debinding) parts obtained in RM process chains do hardly reach densities above 65%. Such green/brown densities are typical for polymer coated powders (e.g. DTM powders), the initial Ni-Cu-P bronze powders (e.g. Directsteel, Electrolux, Mareco powders), WC-Co powders, etc. Such high porosity will normally not allow for post-sintering. Post-sintering tests were done starting from green parts having 65% density. It turned out not to be possible to reach near 100% density, but an asymptotic density of 85% in Ar atmosphere and 95% in vacuum could be reached after some 2 hours [126]. The success of post-sintering not only depends on the initial density, but largely depends on other characteristics, like size and bulkiness of the parts. This is why it is hard to put a lower limit to the required initial density.

In cases where infiltration should be avoided, because it will introduce an undesired low melting infiltration material into the part, while the green density is too low to afford furnace post-sintering at (near) atmospheric pressure, HIPing might be an appropriate post-consolidation process [37]. HIPing however, requires the outer shell to be fully dense, which is not easy to achieve in SLS even when applying different SLS processing parameters for core and skin: see Figure 39 b and Figure 40 b. If the outer part surface cannot be made fully dense during SLS, caning should be applied before HIPing [37] which is generally impossible with parts having complex geometries, as most RP/RM parts are.

5 CONCLUSIONSThis paper has provided a survey of consolidation mechanisms used in laser and powder bed based layered manufacturing, i.e. selective laser sintering/melting (SLS/SLM). It has demonstrated the main physical mechanisms that influence the consolidation and how the material properties and composition play an important

-754-

role. This explains why there are still many restrictions to enlarging the scope of materials that are processed today. Herewith the paper has led the basis for further R&D in coming decades to expand the applicability of SLS/SLM to a much larger range of materials yielding ever better part and material properties.

6 ACKNOWLEDGEMENTS The authors would like to thank all people having contributed to this paper, in particular (1 = CIRP member): Persons having provided input to this paper: M.

Geiger1, M. Schmidt, B. Hon1, J. Meijer1, and the many persons having provided input during the CIRP STC-E meetings

Persons having provided input and contributed to the writing: T. Craeghs, S. Kumar, E. Yasa, J. Van Vaerenbergh, P. Mercelis, B. Vandenbroucke, A. Demmer, T. Glaeser, C. Derichs, K. Dalgarno, M. Badrossamay, R. Schindel, P. Schleiss

All persons having contributed to the CIRP STC-E Working Group on “melt pool phenomena in laser powder processing”.

7 REFERENCES1 3D Systems, 2005, Duraform AF data sheet. 2 Abe, F., Osakada, K., Kitamura, Y., Matsumoto, M., Shiomi, M., 2000, Manufacturing of titanium parts for medical purposes by selective laser melting, Proc. Rapid Prototyping: 288-293. 3 Abe, F., Osakada, K., Shiomi, M., Uematsu, K., Matsumoto, M., 2001, The manufacturing of hard tools from metallic powders by selective laser melting, J. Materials Processing Technology, 111/1-3: 210-213. 4 Agarwala, M.K., Bourell, D.L., Beaman, J., Marcus, H., Barlow, J., 1995, Direct selective laser sintering of metals, Rapid Prototyping J., 1/1: 26-36. 5 Agarwala, M.K., Bourell, D.L., et al., 1993, Synthesis, selective laser sintering and infiltration of high Tc, dual phase Ag - YBaCu3O7−x superconductor composites, Proc. SFF Symp., Austin: 339-349. 6 Aggarangsi, P., Beuth, J.L., 2006, Localized preheating approaches for reducing residual stress in additive manufacturing, Proc. SFF Symp., Austin: 709-720.7 Ajoku, U., Hopkinson, N., Caine, M., 2006, Experimental measurement and finite element modelling of the compressive properties of laser sintered Nylon-12, Materials Science and Engineering, A/428: 211–216. 8 Alexandre, T., Giovanola, J., Vaucher, S., Beffort, O., Vogt, U., 2004, Layered manufacturing of porous ceramic parts from ceramic powders and preceramic polymers, Proc. LANE, Erlangen, 497-504. 9 Antonov, E.N., Bagratashvili, V.N., et al., 2005, Three-dimensional bioactive and biodegradable scaffolds fabricated by surface-selective laser sintering, Advanced Materials, 17/3: 327-333. 10 Aubry, P., Lafaye, Y., Coulon, N., 2007, Analysis of the laser sintering process for direct manufacturing of mould, Proc. Lasers in Manufacturing, Munich: 167-172. 11 Beaman, J.J., Barlow, J.W., Bourell, D.L., Crawford, R.H., Marcus, H.L., McAlea, K.P., 1997, SFF: A new direction in manufacturing, Kluwer Academic Publishers, USA.12 Beer, S.Z., 1972, Liquid metals chemistry and physics, Marcel Dekker, New York. 13 Bertrand, Ph., Yadroitsev, I., Smurov, I., 2004, Prototypage rapide et fabrication directe par laser d’objets

multimatériaux multifonctionnels, 10ième Assisses Europ. de Proto Rapide, Paris, France. 14 Berzins M., Childs T.H.C., Ryder, G.R., 1996, The selective laser sintering of polycarbonate, CIRP Annals, 45/1: 187-190. 15 Bibb, R., Eggbeer, D., Rapid manufacture of removable partial denture frameworks, Rapid Prototyping J., 12/2: 95-99. 16 Bourell, D.L., Evans, R. S., Barrows, S. L., 2005, High temperature infiltration of non-metallic particles produced by selective laser sintering, Materials Science Forum, 475-479: 2861-2866. 17 Bourell, D.L., Wohlert, M., et. al., 2002, Powder densification maps in selective laser sintering, Advanced Engineering Materials, 4/9: 663-669. 18 Bourell, D.L., Beaman, J.J., 2005, Materials issues in rapid prototyping, Proc. VRAP, Leiria: 305-310. 19 Bourell, D.L., Marcus, H.L., Barlow, J.W., Beaman, J.J., 1992, Selective laser sintering of metals and ceramics, Int. J. Powder Metallurgy, 28/4: 369-381. 20 Burning S., 1998, Copper Polyamide, 2nd European SLS users meeting, K.U. Leuven, Belgium. 21 Casalino, G., De Filippis, L.A.C., Ludovico, A., 2005, A technical note on the mechanical and physical characterization of selective laser sintered sand for rapid casting, J. Materials Processing Technology, 166/1: 1-8. 22 Caulfield, B., McHugh, P.E., Lohfeld, S., 2007, Dependence of mechanical properties of polyamide components on build parameters in the SLS process, J. Materials Processing Technology, 182/1-3: 477-488. 23 Chen, X.C., Xie, J.W., Fox, P., 2004, Direct laser remelting of iron with addition of boron, Material Science and Technology, 20: 715-719. 24 Cheng, J., Lao, S., 2005, SLS processing studies of nylon 11 nanocomposites, Proc. SFF Symp., Austin: 141-149.25 Childs, T.H.C., Hauser, C., 2005, Raster scan selective laser melting of the surface layer of a tool steel power bed, Proc. IMechE part B, J. Engineering Manufacture, 219: 379-384. 26 Childs, T.H.C., Hauser, C., Taylor, C.M, Tontovi, A.E., 2000, Simulation and experimental verification of crystalline polymer and direct metal selective laser sintering, Proc. SFF Symp., Austin: 100-109. 27 Childs, T.H.C., Hauser, C., Badrossamay, M., 2005, Selective laser sintering (melting) of stainless and tool steel powders: experiments and modeling, Proc. IMechE part B, J. Engineering Manufacture, 219: 339-357. 28 Childs, T.H.C., Tontowi, A.E., 2001, Selective laser sintering of a crystalline and glass-filled crystalline polymer: experiments and simulations, Proc. IMechE Part B, J. Engineering Manufacture, 215/11: 1481-1495. 29 Chua, C.K., Leong, K.F., Tan, K.H., Wiria, F.E., Cheah, C. M., 2004, Development of tissue scaffolds using selective laser sintering of polyvinylalcohol/hydroxyapatite biocomposite for craniofacial and joint defects, J. Materials Science: Materials in Medicine, 15/10: 1113-1121. 30 Chung, H., Das, S., 2006, Processing and properties of glass bead particulate-filled functionally graded nylon-11 composites produced by selective laser sintering, Materials Science and Engineering A, 437/2: 226-234. 31 Ciardelli, G., Chiono, V., et al., 2005, Blends of poly-(epsilon-caprolactone) and polysaccharides in tissue engineering applications, Biomacromolecules, 6/4: 1961-1977.

-755-

32 Cruz, F., Coole, T., Bocking, C., Simoes, J., 2003, Selective laser sintering of customized medical implants using biocomposite materials, Tech. Vjesn., 10/2: 23-27. 33 Cruz, F., Simoes, J., Coole, T., Bocking, C., 2005, Direct manufacture of hydroxyapatite based bone implants by selective laser sintering, Proc. VRAP, Leiria: 119-126. 34 Dalgarno, K.W., Wood, D.J., et al., 2005, Mechanical properties and biological responses of bioactive glass ceramic processed using indirect SLS, Proc. SFF Symp., Austin: 132-140. 35 Das, S., 2003, Physical aspects of process control in selective laser sintering of metals, Advanced Engineering Materials, 5: 701-711. 36 Das, S., Harlan, N., et.al., 1998, Direct SLS processing for production of cermet composite turbine sealing components, Materials and Manufacturing Processes, 13/2: 1-17. 37 Das, S., Wohlert, M., Beaman, J.J., Bourell, D.L., 1998, Processing of Titanium Net Shapes by SLS/HIP, Proc. SFF Symp., Austin: 469 – 477. 38 Davis, S.H., 1980, Moving contact lines and rivulet instabilities, Part 1, The static rivulet, J. Fluid Mechanics, 98: 225-242. 39 Deckard, L., Dennis Claar, T., 1993, Fabrication of ceramic and metal matrix composites from selective laser sintered performs, Proc. SFF Symp., Austin: 215-222. 40 Dickens, D.E., Lee, B.L., Taylor, G.A., Magistro, A.J., Ng, H., McAlea, K.P., Forderhase, P.F., 1997, US patent 5648450: Sinterable semi-crystalline powder and near-fully dense article formed therein. 41 Dimov, S.S., Pham, D.T., et al., 2001, Rapid tooling applications of the selective laser sintering process, Assembly Automation, 21/4: 296-302. 42 Dittrich, T., Sieber I., Henrion, W., Rauscher, S., Wanderka, N., Rappich, J., 1996, Selective laser induced melting of ultrathin nonporous silicon layers, Applied Physics A: Materials Sci. and Processing, 63/5: 467-470. 43 Dolinsek, S., Panjan, P., Syvanen, T., Ramovs, J., 2006, Application of rapid tooling for aluminium die casting, Proc. Euro-u Rapid, Frankfurt. 44 Duck, J., Niebling, F., Neesse, T., Otto, A., 2004, Infiltration as post-processing of laser sintered metal parts, Powder Technology, 145/1: 62-68. 45 EOS – website: http://www.eos-gmbh.de 46 EOS, Alumide data sheet 47 Evans, R.S., Bourell, D.L., Beaman, J.J., Campbell, M.I., 2005, Rapid manufacturing of silicon carbide composites, Rapid Prototyping J., 11/1: 37-40. 48 Fan, K.M., Cheung, W.L., Gibson, I., 2005, Movement of powder bed material during the selective laser sintering of bisphenol-a-polycarbonatem, Rapid Prototyping J., 11/4: 188-198. 49 Feenstra, F., Holmer, B., Pohl, H., Tromans, G., Moos, N., Mieritz, B., 2002, RP, RT, RM trends and developments / research survey, Technical report, RAPTIA project. 50 Fischer, H., Wilkes, J., Bergmann, C., Kuhl, I., Meiners, W., Wissenbach, K., Poprawe, R., Telle, R., 2006, Bone substitute implants made of TCP/Glass composites using selective laser melting technique, RWTH Aachen University, Fraunhofer Institute for Laser Technology ILT, Ber DKG 83:57-60. 51 Fischer, P., Leber, H., Romano, V., Weber, H.P., Karapatis, N.P., André, C., Glardon, R., 2004, Microstructure of near-infrared pulsed laser sintered titanium samples, Applied Physics A, 78/8: 1219-1227.

52 Fischer, P., Romano, V., Weber, H.P., Karapatis, N. P., Boillat, E., Glardon R., 2003, Sintering of commercially pure titanium powder with a Nd:YAG laser source, Acta Materialia, 51: 1651-1662. 53 Fisher P., Karapatis, N., Romano, V., Weber, H.P., 2002, A model for the interaction of near infrared pulsed laser with metal powders in selective laser sintering, Applied Physics A, 74/4: 467-474. 54 Friedel, T., Travitzky, N., Niebling, F., Scheffler, M., Greil, P., 2005, Fabrication of polymer derived ceramic parts by selective laser curing, J. European Ceramic Society, 25/2-3: 193-197. 55 Gaard, A., Krakhmalev, P., Bergstrom, J., 2006, Microstructural characterization and wear behaviour of (Fe,Ni)-TiC MMC prepared by DMLS, J. Alloys and Compounds, 421/1-2: 166-171. 56 Gahler, A., Heinrich, J.G., 2006, Direct laser sintering of Al2O3–SiO2 dental ceramic components by layer-wise slurry deposition, J. American Ceramic Society, 89/10: 3076-3080. 57 Gauckler, L.J., 2000, Ingenieurkeramik 2 - Herstellung von Keramik, Vol. 2. ETH-Zurich, Zurich. 58 Ghany, A.K., Moustafa, S.F., 2006, Comparison between the products of four RPM systems for metals, Rapid Prototyping J., 12/2: 86-94. 59 Gill, T., Hon, B., 2002, Selective laser sintering of SiC/PA matrix composites, Proc. SFF Symp., Austin: 538-545.60 Glardon, R., Karapatis, N., Romano, V., 2001, Influence of Nd:YAG parameters on the selective laser sintering of metallic powders, CIRP Annals, 50/1: 133-136. 61 Goodridge, R.D., Wood, D.J., Ohtsuki, C., Dalgarno, K. W., 2007, Biological evaluation of an apatite-mullite glass-ceramic produced via selective laser sintering, Acta Biomaterialia, 3/2: 221-231. 62 Greco, A., Maffezzoli, A., 2003, Polymer melting and polymer powder sintering by thermal analysis, J. Thermal Analysis and Calorimetry, 72/3: 1167-1174. 63 Gu, D., Shen, Y., 2006, WC-Co particulate reinforcing Cu matrix composites produced by direct laser sintering. Materials Letters, 60/29-30: 3664-3668. 64 Gu, D., Shen, Y., Fang, S., Xiao, J., 2007, Metallurgical mechanisms in direct laser sintering of Cu-CuSn-CuP mixed powder, J. Alloys and Compounds, 438: 184-189.65 Gu, D., Shen, Y., Zhao, L., Xiao, J., Wu, P., Zhu, Y., 2007, Effect of rare earthoxide addition on microstructures of ultra-fine WC-Co particulate reinforced Cu matrix composites prepared by direct laser sintering, Materials Science and Engineering A, 445-446: 316-322. 66 Gusarov, A.V., 2003, Mechanisms of selective laser sintering and heat transfert in Ti powder, Rapid Prototyping J., 9/5: 314-326. 67 Gusarov, A.V., Laoui, T., Froyen, L., Titov, V., Tolochko, N., 2001, Numerical simulation of laser solid state sintering of loose titanium powder, Proc. 10th Europ. Conf. on Rapid Prototyping and Manufacturing, Paris. 68 Han, Z., Cao, W.B., et al., 2004, Progress on rapid prototyping of ceramic parts by selective laser sintering, J. Inorganic Materials, 19/4: 705-713. 69 Hao, L., Savalani, M.M., Harris, R.A., 2005, Layer manufacturing of polymer/bioceramic implants for bone replacement and tissue growth, Proc. VRAP, Leiria: 127-132.70 Hao, L., Savalani, M.M., Zhang, Y., et al., 2006, Effect of material morphology and processing conditions on the characteristics of hydroxyapatite and high-density

-756-

polyethylene biocomposite by selective laser sintering. Proc. IMechE Part L, J. Materials Design and Applications, 220/L3: 125-137. 71 Harlan, N., Bourell, D.L., Beaman, J., 1998, Titanium casting molds via selective laser sintering, Proc. SFF Symp., Austin: 597-604. 72 Harlan, N., Park, S., Bourell, D.L., Beaman, J., 1999, Selective laser sintering of zirconia with micro-scale features, Proc. SFF Symp., Austin: 297-302. 73 Hauser, C., 2003, Selective laser sintering of a stainless steel powder, PhD thesis, Univ. of Leeds, UK. 74 Hauser, C., Childs, T.H.C., Dalgarno, K.W., Eane, R. B., 1999, Atmospheric control during direct selective laser sintering of stainless steel 314s powder. Proc. SFF Symp., Austin: 265-272. 75 Ho, H.C.H., Cheung, W.L., Gibson, I., 2002, Effects of graphite powder on the laser sintering behaviour of polycarbonate. Rapid Prototyping J., 8/4: 233-242. 76 Hon, K.K.B., Gill, T.J., 2003, Selective laser sintering of SiC/polyamide composites, CIRP Annals, 52/1: 173-176.77 Imai, Y., Kyogoku, H., Shiraishi, K., 2005, Laser sintering of stainless steel using resin powder, Proc. SFF Symp., Austin: 254-260. 78 Ji, L., and Jiang, Y., 2006, Dielectric and microstructure modification of Ta2O5 ceramics by laser sintering, Materials Letters, 60/1: 86-89. 79 Ji, L., and Jiang, Y., 2006, Laser sintering of transparent Ta2O5 dielectric ceramics, Materials Letters, 60/12: 1502-1504. 80 Kamitani, T., Yamada, O., Marutani, Y., 2000, Selective laser sintering with heat of formation by using reactive materials, Proc. SPIE, 4088: 299–302. 81 Kandis, M., Bergman, T.L., 2000, A simulation-based correlation of the density and thermal conductivity of objects produced by laser sintering of polymer powders, J. Manuf. Sci. and Eng.,122/3: 439-444. 82 Karapatis, P., 2002, A sub-process approach of selective laser sintering, PhD thesis, Ecole Polytechnique Federale de Lausanne, Switzerland. 83 Keene, B.J., Mills, K.C., Brooks, R.F., 1985, Surface properties of metals and their effect on weldability, Materials Science and Technology, 1: 568–571. 84 Kim, J., Creasy, T. S., 2004, Selective laser sintering characteristics of nylon 6 / clay - reinforced nanocomposite, Polymer Testing, 23: 629-636. 85 Kitamura, Y., Santos, E., Abe, F., Osakada, K., Shiomi, M., 2000, Processing of titanium by selective laser melting for medical purposes, Proc. Rapid Prototyping & Manufacturing, Malmoe, Sweden. 86 Klocke, F., Ader, C., 2004, Layer manufacturing of ceramic materials, Proc. LANE, Erlangen: 505-510. 87 Klocke, F., Wagner, C., 2003, Coalescence of two metallic particles as base mechanism of selective laser sintering, CIRP Annals, 52/1: 177-184. 88 Klocke, F., Wirtz, H., 1997, Selective laser sintering of ceramics, Proc. LANE, Erlangen: 589-596. 89 Kolossov, S., Boillat, E., Glardon, R., Fischer, P., Locher, M., 2004, 3D FE Simulation for temperature evolution in the selective laser sintering process, Int. J. Machine Tools and Manufacture, 44:117-123. 90 Koo, J.H., Lao, S., et al., 2006, Polyamide nanocomposites for selective laser sintering, Proc. SFF Symp., Austin: 392-409.

91 Koo, J.H., Pilato, L., et al., 2005, Innovative selective laser sintering rapid manufacturing using nanotechnology, Proc. SFF Symp., Austin: 98-108. 92 Kralchevsky, P.A., Denkov, N.D., 2001, Capillary forces and structuring in layers of colloid particles, Current Opinion in Colloid & Interface Sci., 6: 383-401. 93 Kruth, J.-P., Froyen, L., Rombouts, M., Van Vaerenbergh, J., Mercelis, P., 2003, New ferro powder for selective laser sintering of dense parts, CIRP Annals, 52/1: 139-142. 94 Kruth, J.-P., Kumar, S., Van Vaerenbergh, J., 2005, Study of laser-sinterability of ferro-based powders, Rapid Prototyping J., 11/5: 287-292. 95 Kruth, J.-P., 1991, Material incress manufacturing by rapid prototyping techniques, CIRP Annals, 40/2: 603-614. 96 Kruth, J.-P., Froyen, L., Rombouts, M., Van Vaerenbergh, J., Mercelis, P., 2003, New ferro powder for selective laser sintering of dense parts, CIRP Annals, 52/1: 139–142. 97 Kruth, J.-P., Froyen, L., Van Vaerenbergh, J., Mercelis, P., Rombouts, M., Lauwers, B., 2004, Selective laser melting of iron based powder, J. Materials Processing Technology, 149/1 - 3: 616 – 622. 98 Kruth, J.-P., Leu, M. C., Nakagawa, T., 1998, Progress in additive manufacturing and rapid prototyping. CIRP Annals, 47/2: 525-540. 99 Kruth, J.-P., Mercelis, P., Van Vaerenbergh, J., Froyen, L., Rombouts, M., 2005, Binding mechanisms in selective laser sintering and selective laser melting, Rapid Prototyping J., 55/1: 26-36. 100 Kruth, J.-P., Peeters, P., Smolderen, Th., Bonse, J., Laoui, T., Froyen, L., 1998, Comparison between CO2 and Nd:YAG lasers for use with selective laser sintering of steel-copper powders, Revue Intern. de CFAO et d'informatique graphique, 13/4-5-6: 95-112. 101 Kruth, J.-P., Van der Scheuren B., Bonse, J.E., Morren, B., 1996, Basic powder metallurgical aspects in selective metal powder sintering, CIRP Annals, 45/1: 183-186.102 Kruth, J.-P., Van Elsen, M., 2005, Manufacturing of light weight polymer structures using SLS, Proc. PMI, Gent, CD-Rom. 103 Kruth, J.-P., Vandenbroucke B., Van Vaerenbergh, J., Mercelis P., 2005, Benchmarking of different SLS/SLM processes as rapid manufacturing techniques, Proc. PMI, Gent, CD-Rom. 104 Kruth, J.-P., Vandenbroucke, B., Van Vaerenbergh, J., Naert, I., 2005, Digital manufacturing of biocompatible metal frameworks for complex dental prostheses by means of SLS/SLM, Proc. VRAP, Leiria: 139-146. 105 Kruth, J.-P., Wang, X., Laoui, T., Froyen, L., 2003, Lasers and materials in selective laser sintering, Assembly Automation, 23/4: 357-371. 106 Kumar, S., Kruth, J.-P., 2007, Effect of bronze infiltration on laser sintered metallic parts, Materials and Design, 28/2: 400-407. 107 Kumetsov, M.V., Parkin, I.P., et al., 2006, Advanced ways and experimental methods in self-propagating high temperature synthesis (SHS) of inorganic materials, Materials Science Forum, 518: 181-188. 108 Kun, X., 2007, Indirect Selective Laser Sintering of Apatite-Wollostonite, PhD Thesis, Univ. of Leeds, UK. 109 Kyogoku, H., Ramos, J.A., Bourell, D.L., 2002, Laser melting of Ti-Ni shape memory alloy, Proc. SFF Symp., Austin: 668-675.

-757-

110 Lancaster, J.F., 1986, The Physics of Welding, 2nd

edition, The International Institute of Welding, Pergamon Press, Oxford. 111 Lanzetta, M., Santochi, M., 2006, Liquid-phase infiltration of thermal sintered skeletons by low-temperature gold eutectic alloys, CIRP Annals, 55/1: 213-216.112 Laoui, T., Froyen, L., Kruth, J.P., 2000, Effect of mechanical alloying on Selective Laser Sintering of WC-9Co powder, Powder Metallurgy, 42/3: 203-205. 113 Laoui, T., Santos, E., Osakada, K., Shiomi, M., Morita, M., Shaik, S.K., Tolochko, N.K., Abe, F., 2004, Properties of titanium implant models made by laser processing, Proc. LANE, Erlangen: 485-496. 114 Laoui, T., Wang, X., Childs, T.H.C., Kruth, J.P., Froyen, L., 2000, Laser penetration in a powder bed during selective laser sintering of metal powders: simulations versus experiments, Proc. SFF Symp., Austin: 7-9.115 Lee, G., Barlow, J.W., Fox, W.C., Aufdermorte, T.B., 1996, Biocompatibility of SLS-formed calcium phosphate implants, Proc. SFF Symp., Austin: 15-21. 116 Lee, I.S., 1998, Rapid full densification of alumina-glass composites fabricated by a selective laser sintering process, J. Materials Science Letters, 17/22: 1907-1911. 117 Lenk, R., Nagy, A., Techel, A., 2003, Material development for laser sintering of high temperature strength silicon carbide with extern hardness. Proc. Euro-uRapid, Frankfurt: B/4, 1-4. 118 Leong, C.C., Lu, L., Fu, J.Y.H., Wong, Y. S., 2002, In-situ formation of copper matrix composites by laser sintering, Materials Science & Engineering A, 338: 81-88. 119 Levy, G.N., 2006, Total quality management for rapid manufacturing, Proc. Euro-uRapid Frankfurt: A5-1. 120 Levy, G.N., Boehler, P., 2005, Controlled local properties in the same part with sintaflex: a new elastomer powder for the sls process, Proc. SFF Symp., Austin: 197-207.121 Levy, G.N., Schindel, R., Kruth, J.-P., 2003, Rapid manufacturing and rapid tooling with layer manufacturing technologies: state of the art and future perspectives, CIRP Annals, 52/2: 589-609. 122 Liu, H.X., Yan, Y.N., Wang, X.H., Cheng, J., Pan, Y.Q., Lin, F., Xiong, Z., 2005, The direct cell controlled assembly technology on the base of RP, Proc. VRAP, Leiria: 91-96.123 Liu, J., Shi, Y., Lu, Z., Xu., Y., Huang, S., 2006, Rapid manufacturing metal parts by laser sintering admixture of epoxy resin/iron powders, Advanced Materials Engineering, 8/10: 988-994. 124 Loschau, W., Lenk, R., Scharek, S., Teichgraber, M., Nowotny, S., Richter, C., 2000, Prototyping of complex-shaped parts and tools of Si/SiC-ceramics by selective laser sintering, Industrial Ceramics, 20/2: 95-96. 125 Lu, L., Fuh, J.Y.H., Chen, Z.D., Leong, C.C., Wong, Y.S., 1999, In situ formation of TiC composite using selective laser melting, Materials Research Bulletin, 35: 1555-1561.126 Lu, L., Fuh, J., Wong, Y.-S., 2001, Laser-induced materials and processes for rapid prototyping, Kluwer Academic Publishers, Massachusetts 02061, USA. 127 Lu, S., Fujii, H., Nogi, K., 2004, Marangoni convection and weld shape variations in Ar–O2 and Ar–CO2 shielded GTA welding, Materials Science and Engineering A, 380/1-2: 290-297. 128 Lu, S., Fujii, H., Nogi, K., 2004, Sensitivity of Marangoni convection and weld shape variations to

welding parameters in O2–Ar shielded GTA welding, Scripta Materialia, 51/3: 271-277. 129 Macedo, Z.S., Hernandes, A.C., 2002, Laser sintering of bismuth germanate (Bi4Ge3O12) ceramics, J. American Ceramic Society, 85/7: 1870-1872. 130 Macedo, Z.S., Hernandes, A.C., 2002, Laser sintering of Bi4Ti3O12 ferroelectric ceramics, Materials Letters, 55: 217-220. 131 Macedo, Z.S., Hernandes, A.C., 2005, A quantative analysis of the laser sintering of bismuth titanate ceramics, Materials Letters, 59/27: 3456-3461. 132 Maeda, K. Childs, T. H. C., 2004, Laser sintering (SLS) of hard metal powders for abrasion resistant coatings, J. Materials Proc. Tech., 149/1-3: 609-615. 133 Mazzoli, A., Moriconi, A., Pauri, M.G., 2007, Characterization of an aluminium-filled polyamide powder for applications in selective laser sintering, Materials & Design, 28/3: 993-1000. 134 Mercelis, P., Kruth, J.-P., 2006, Residual stresses in selective laser sintering and selective laser melting, Rapid Prototyping J., 12/5: 254-265. 135 Mercelis, P., 2007, Control of selective laser sintering and selective laser melting, PhD thesis, Katholieke Universiteit Leuven, Belgium. 136 Mercelis, P., Van Vaerenbergh, J., Kruth, J.-P., 2007, Feedback control of Selective laser melting, Proc. ISEM, Pittsburgh, USA: 421-426. 137 Mills, K.C., Keene, B.J., Brooks, R.F., Shirali, A., 1998, Marangoni effects in welding, Phil. Trans. R. Soc. Lond. A, 356: 911-925.138 Mumtaz, K.A., Hopkinson, N., Erasenthiran, P., 2006, High density selective laser melting of waspaloy, Proc. SFF Symp., Austin: 220-232. 139 National Council for Advanced Manufacturing NACFAM, Report on Industry Views Towards: Categories of Innovative and Potentially Disruptive Advanced Manufacturing Technologies, April 2005, Washington DC. 140 Nelson, J.C., Vail, N.K., et al., 1995, Selective laser sintering of polymer-coated silicon carbide powders, Industrial & Eng. Chemistry Res., 34/5: 1641-1651. 141 Nelson, J.C., Xue, S., et al., 1993, Model of the selective laser sintering of bisphenol-a polycarbonate, Ind. Eng. Chem. Res., 32: 2305-2317. 142 Nemchinsky, V.A., 1997, The role of thermocapillary instability in heat transfer in a liquid metal pool, J. Heat and Mass Transfer, 40/4: 881-891. 143 Niino, T. Yamada, H., 2005, Transparentization of SLS processed SMMS copolymer parts by infiltrating a thermosetting epoxy resin with tuned refractive index, Proc. SFF Symp., Austin, 208-216. 144 O’Neill, W., 1998, Investigation of short pulse Nd:YAG laser interaction with stainless steel powder beds, Proc. SFF Symp., Austin, 147-153. 145 Over, C., 2003, Generative Fertigung von Bauteilen aus Werkzeugstahl X38CrMoV5-1 und Titan TiAl6V4 mit Selective Laser Melting, PhD Thesis, RWTH Aachen. 146 Over, C., Meiners, W., Wissenbach, K., Hutfless, J., Lindeman, M., 2003, Rapid Manufacturing of metal parts and tools using laser melting, Proc. WLT-Conference on Lasers in Manufacturing, Munich: 265-269. 147 Over, C., Meiners, W., Wissenbach, K., Lindemann, M., Hammann, G., 2001, Selective laser melting: a new approach for the direct manufacturing of metal parts and tools, Proc. LANE, Erlangen: 391-398. 148 Partee, B., Hollister, S.J., Das, S., 2006, Selective laser sintering process optimization for layered

-758-

manufacturing of CAPA® 6501 polycapralactone bone tissue engineering scaffolds, J. Manufacturing Science and Engineering-Transactions ASME, 128/2: 531-540. 149 Pogson, S.R., Fox, P., et al., 2003, The production of copper parts using DMLR, Rapid Prototyping J., 9/5: 334-343.150 Poprawe, R., 2004, Development of new beam sources - fiber, slab, disc and rod. Laser Assisted Net Shape Engineering 4, Proc. LANE, Erlangen: 47-54. 151 Rechtenwald, T., Niebling, F., Pohle, D., Esser, G., 2004, An investigation into laser sintering of PEEK, Proc. LANE, Erlangen: 485-496. 152 Rechtenwald, T., Pohle, D., Otto, A., Schmidt, M., 2007, Dimensional accuracy and surface finish of laser sintered polyetheretherketone parts, Proc. Laser in Manufacturing, Munich. 153 Regenfuss, P., Hartwig, L., et al., 2005, Industrial freeform generation of microtools by laser micro sintering, Rapid Prototyping J., 11/1: 23-26. 154 Regenfuss, P., Streek, A., et al.,2006, Principles of laser micro sintering, Proc. SFF Symp., Austin: 740-753. 155 Rimell, J.T., Marquis, P.M., 2000, Selective laser sintering of ultra high molecular weight polyethylene for clinical applications, J. Biomedical Materials Research, 53/4: 414-420. 156 Rombouts, M., 2006, Selective laser sintering/melting of iron-based powders, PhD thesis, Katholieke Universiteit Leuven, Belgium. 157 Rombouts, M., Froyen, L., Kruth, J.P., Van Vaerenbergh, J., Mercelis, P., 2004, Production and properties of dense iron based parts produced by laser melting with plasma formation, Proc. Powder Metallurgy World Congress, 5: 115-121. 158 Rombouts, M., Kruth, J.-P., Froyen, L. Mercelis, P., 2006, Fundamentals of selective laser melting of alloyed steel powders, CIRP Annals, 55/1: 197-192. 159 Rople, S.I., Tsarevskii, B.V., et al., 1975, Combined influence of oxygen and sulphur on the surface tension of iron, Izv. AN. SSSR Metally, 4: 54-60. 160 Sadeghian, Z., Heinrich, J.G., Moztarzadeh, F., 2004, Direct laser sintering of hydroxyapatite implants by layer-wise slurry deposition, Ceramic Forum International, 81/12: E39-E43.161 Salmang, H., Scholze, H., 1982, Keramik Teil1: Allgemeine Grundlagen und wichtige Eigenschaften. 6th

Ed. Springer, Berlin. 162 Salmoria, G., et al., 2007, Rapid manufacturing of PA/HDPE blend specimens by selective laser sintering: microstructural characterization, Polymer Testing, 26: 361-368.163 Santos, E.C., Abe, F., Kitamura, Y., Osakada, K., Shiomi, M., 2002, Mechanical properties of pure titanium models processed by selective laser melting. Proc. SFF Symp., Austin: 180-186. 164 Santos, E.C., Shiomi, M., Osakada, K., Laoui, T., 2006, Rapid manufacturing of metal components by laser forming, Int. J. Machine Tools & Manuf., 46: 1459-1468. 165 Savalani, M.M., Hao, L., Harris, R.A., 2006, Evaluation of CO2 and Nd:YAG lasers for the selective laser sintering of HAPEX®, Proc. IMechE Part B, J. Engineering Manufacture, 220/2: 171-182. 166 Savalani, M.M., 2006, Selective laser sintering of hydroxyapatite – polyamide composites, PhD thesis, Loughbourough University, Leeds, UK. 167 Savalani, M.M., Hao, L., Zhang, Y., Tanner, K.E., Harris, R.A., 2007, The effects and interactions of fabrication parameters on the properties of selective laser

sintered hydroxyapatite polyamide composites, Proc. IMechE Part B, J. of Engineering Manufacture, submitted. 168 Saxhlos, E., Reis, N., Ainsley, C., Derby, B., Czernuszka, J.T., 2003, Novel collagen scaffolds with predefined internal morphology made by SFF, Biomaterials, 24: 1487-1497. 169 Schiaffino, S., Sonin, A.A., 1997, Formation and stability of liquid and molten beads on a solid surface, J. Fluid Mechanics, 343: 95-110. 170 Schindel, R., Levy, G., Zinn, M., Lampert, C., 2004, Biomaterials are going 3D with layer manufacturing technologies (LM) – direct and indirect approaches, 7th

Essen Symp. on Biomaterials and Biomechanics, University Duisburg-Essen. 171 Sercombe, T.B., Schaffer, G.B., 2003, Rapid manufacturing of aluminium components, Science, 301: 1225-1227.172 Sercombe, T.B., Schaffer, G.B., 2003, The production of aluminium SLS prototypes via infiltration, Proc. VRAP: 297-301. 173 Sercombe, T.B., Schaffer, G.B., 2004, On the role of magnesium and nitrogen in the infiltration of aluminium by aluminium for rapid prototyping applications, Acta Materialia, 52/10: 3019-3025. 174 Shi, Y.S., Li, Z.C., 2004, Development of polymer alloy of polysterene (PS) and polyamide (PA) for building functional part based on selective laser sintering, Proc. IMechE Part L, J. Materials Design and Applications, 218/L4: 299-306. 175 Shi, Y.S., Li, Z.C., 2004, Effect of the properties of polymer materials on the quality of selective laser sintering parts, Proc. IMechE Part L, J. Materials Design and Applications, 218/L3: 247-252. 176 Shi, Y., Chen, S., Lu, X., Huang, S., 2003, An investigation into adopting SHS technology in SLS metallic material, Proc. VRAP, Leiria: 289-296. 177 Shishkovsky, I.V., Kuznetsov, M.V., 2004, Laser-induced combustion synthesis of 3D functional materials: Computer-Aided Design, J. Materials Chemistry, 14/23: 3444-3448.178 Shishkovsky, I.V., Tarasov, E.Y., Zhuravel, L.V., Petrov, A.L., 2001, Selective laser sintering in the powder synthesis of a biocomposite based on titanium nickelide and hydroxyapatite, Tech. Physics Letters, 27/3: 211-213. 179 Shishkovsky, I.V., Tarasova, E., Petrov, A., 2001, Synthesis of biocomposite on the base of NiTi with hydroxyapatite by selective laser sintering, Proc. LANE, Erlangen: 459-464. 180 Simchi, A. Pohl, H., 2004, Direct laser sintering of iron-graphite powder mixture, Materials Science & Engineering A, 383/2: 191-200. 181 Simchi, A., Petzoldt F., Pohl, H., 2003, On the development of direct metal laser sintering for rapid tooling, J. Materials Proc. Technology, 141/3: 319-328. 182 Slocombe, A., Li, L., 2001, Selective laser sintering of TiC-Al2O3 composite with self-propagating high temperature synthesis, J. Materials Processing Technology, 118: 173-178. 183 Song, J.L., Li, Y.T., Deng, Q.L., Hu, D.J, 2007, Rapid prototyping manufacturing of silica sand patterns based on selective laser sintering, J. Materials Processing Technology, 187-188/12: 614 - 618. 184 Song, Y.-A., 1997, Experimental study of the basic process mechanism for direct selective laser sintering of low-melting metallic powder, CIRP Annals, 46/1:127-130. 185 Starly, B., Lau, A., Sun, W., Lau, W., Bradbury, T., 2004, Biomimetic design and fabrication of interior

-759-

architecture of tissue scaffolds using SFF, Proc. SFF Symp., Austin: 742-753. 186 Stevinson, B.Y., Bourell, D.L., Beaman J.J., Support-free infiltration of selective laser sintered silicon carbide performs, Proc. SFF Symp., Austin: 359-365. 187 Streek, A., Regenfuss, P., Ullmann, F., Hartwig, L., Ebert, R., Exner, H., 2006, Processing of silicon carbide by laser micro sintering, Proc. SFF Symp., Austin: 349-358.188 Strietzel, Production of metallic frames using the BEGO Medifacturing process, J. Dental Technology, 50: 37-44.189 Subramanian, P.K., Zong, G., Marcus, H.L., 1992, Selective laser sintering and reactive sintering of ceramic composites, Proc. SFF Symp., Austin: 63-71. 190 Sun, M.M., 1991, Physical modeling of the selective laser sintering process, Diss. University of Texas at Austin, Austin, USA. 191 Sun, M.M., Beaman, J.J., 1991, A three dimensional model for selective laser sintering, Proc. SFF Symp., Austin: 102-109. 192 Tan, K.H., Chua, C.K., Leong, K.F., et al., 2005, SLS of biocompatible polymers for applications in tissue engineering, Bio-Medical Materials and Engineering, 15: 113-124.193 Tang, Y., Fu, J.Y.H., Loh, H.T., Wong, Y.S., Lu, L., 2003, Direct laser sintering of a silica sand, Materials & Design, 24/8, 623-629.194 Tolochko, N.K., Artushkevich, A. S., Laoui, T., Kruth, J.-P., Froyen, L., 2003, Medical testing of dental implants fabricated by laser processing of titaniumpowders, Proc. VRAP, Leira: 629-632. 195 Tolochko, N.K., Mozzharov, S.E., Sobolenko, N.V., et al., 1995, Main relationships governing laser sintering of loose single-component metallic powders, J. Advanced Materials, 2/2: 151-157. 196 Tolochko, N.K., Laoui, T., et al., 2000, Absorptance of powder materials suitable for laser sintering, Rapid Prototyping J., 6/3: 155-160. 197 Tontowi, A.E. Childs, T.H.C., 2001, Density prediction of crystalline polymer sintered parts at various powder bed temperatures, Rapid Prototyping J., 7/3: 180-184.198 Tosoh, 2005, Tosoh zirconia powder “E” grades: new improved zirconia powder. Company paper of Tosoh, Tokio. 199 Uzunsoy, D., Chang, I.T.H., 2005, The effect of infiltrant choice on the microstructure and mechanical properties of Rapidsteel 2.0, Materials Letters, 59/22: 2812-2817.200 Vail, N.K., Balasubramanian, B., Barlow, K.W., Marcus, H.L., 1996, A thermal model of polymer degradation during selective laser sintering of polymer-coated ceramic powders, Rapid prototyping J., 1996, 2/3: 24-40 . 201 Vail, N.K., Barlow, J.W., Beaman, J.J. Marcus, H.L., Bourell, D.L., 1994, Development of a Poly (methyl methacrylate – co – n-butyl methacrylate) Copolymer Binder System, J. Appl. Polym. Sci., 52: 789-812. 202 Vail, N.K., Swain, D., Fox, W.C., Aufdemorte, T.B., Lee, G., Barlow, J.W., 1998, Materials for biomedical applications, Proc. SFF Symp., Austin: 621-628. 203 Van Bael, S., 2006, Ontwerp en productie van botscaffolds via selectief laser sinteren/smelten, Master Thesis, Groep T Engineering School, Belgium.

204 Van Der Schueren, B., 1996, Basic contributions to the development of the selective metal powder sintering process. PhD thesis, K.U.Leuven, Eng. Faculty, Belgium. 205 Vandenbroucke, B., Kruth, J.-P., 2006, Selective laser melting of biocompatible metals for rapid manufacturing of medical parts, Proc. SFF Symp., Austin: 148-159.206 Vaucher, S., Corenno-Morelli, E., et al., 2003, Selective laser sintering of aluminium and titanium-based composites: processing and characterization. Physica Status Solidi A- Applied Research, 199/3: R11-R13. 207 Vaucher, S., Paraschivescu, D., Andre, C., Beffort O., 2002, Selective laser sintering of aluminium-silicon carbide metal matrix composites, Proc. Materials Week, Munich, Germany. 208 Voet, A., Dehaes, J., Mingneau, J., Kruth, J.-P., Van Vaerenbergh, J., 2005, Study of the wear behaviour of conventional and rapid tooling mould materials. Proc. PMI, Gent, CD-ROM. 209 Wagner, T., Hofer, T., et al., 2005, Laser sintering of high temperature resistant polymers with carbon black additives, International Polymer Proc., 19/4: 395-401. 210 Wang, X.C., Laoui, T., et al, 2002, Direct laser sintering of hard metal powders: experimental study and simulation, Int. J. Advanced Manufacturing Technology, 19/6: 411-417. 211 Wang, Y., Shi, Y., Huang, S., 2005, Selective laser sintering of polyamide-recotrite composite, Proc. IMechE Part L, J. Materials Design and Applic., 219/L1: 11-15. 212 Wendel, B., Dallner, C., Schmachtenberg, E., 2007, New developments in selective laser sintering of polymers, Proc. LANE, Erlangen: 323-331. 213 Wilkes, J., Wissenbach, K., 2006, Rapid manufacturing of ceramic components for medical and technical applications via selective laser melting, Proc. Euro-uRapid, Frankfurt: A4-1. 214 Williams, J.M., Adewunmi, A., Schek R.M., Flanagan, C.L., Krebsbach, P.H., Feinberg, S.E., Hollister, S.J., Das, S., 2005, Bone tissue engineering using PCL scaffolds fabricated via SLS, Biomaterials 26: 4817-4827. 215 Wiria, F.E., Leong, K.F., Chua, C.K., Liu, Y., 2007, Poly-e-capralactone/hydroxyapatite for tissue engineering scaffold fabrication by selective laser sintering, Acta Biomaterialia, 3/1:1-12. 216 Wohlers, T.T., 2005, Wohlers Report 2006: Rapid Prototyping & Manufacturing State of the Industry; Annual Worldwide Progress Report. Wohlers Associates. 217 Wu, W.Z., Yan, M.G., 2002, Development of polymer coated metallic powder for selective laser sintering (SLS) process, J. Adv. Materials, 34/2: 25-28. 218 Zarringhalam, H., Hopkinson, N., Kamperman, N.F., de Vlieger, J.J., 2006, Effects of processing on microstructure and properties of SLS Nylon 12, Materials Science and Engineering A, 435–436: 172-180. 219 Zheng, H., Zhang, J., et al., 2006, Effect of core-shell composite particles on the sintering behaviour and properties of nano-Al2O3/polysterene composites prepared by SLS, Materials Letters, 60/9-10: 1219-1223. 220 Zhu, H.H., Lu, L., Fu, J.Y.H., Wu C.C., 2006, Effect of braze flux on direct laser sintering of Cu-based metal powder, Materials & Design, 27/2: 166-170.