WO2008013483A1

WO2008013483A1 - Method and device for producing three-dimensional objects

Info

Publication number: WO2008013483A1
Application number: PCT/SE2006/000917
Authority: WO
Inventors: Morgan Larsson; Anders Snis
Original assignee: Arcam Ab
Priority date: 2006-07-27
Filing date: 2006-07-27
Publication date: 2008-01-31
Also published as: KR20090049576A; CN101479064A; JP2009544501A; JP5108884B2; US8187521B2; KR101271243B1; EP2049289A4; EP2049289A1; EP2049289B1; US20100007062A1; US20120211926A1; CN101479064B

Abstract

The invention concerns a method for producing three-dimensional objects (3) layer by layer using a powdery material (5) which can be solidified by irradiating it with a high-energy beam. The inventive method comprises a pre- heating step with the general purpose of pre-heating the powdery material (5) in a homogeneous manner, followed by a solidifying step with the general purpose of fusing together the powdery material, wherein the pre-heating step comprises the sub-step of scanning a pre-heating powder layer area (10) by scanning the beam along paths (P1.1 - P5.20) distributed over the pre-heating powder layer area (10), wherein consecutively scanned paths (PM. N, P(M+1 ).N) are separated by, at least, a minimum security distance (ΔY), said minimum security distance (ΔY) being adapted to prevent undesirable summation effects in the pre-heating powder layer area (10) from said consecutively scanned paths. The invention also concerns a device adapted to be operated by the inventive method.

Description

TITLE

Method and device for producing three-dimensional objects.

TECHNICAL FIELD

This invention relates to a method and device for producing three- dimensional objects layer by layer using a powdery material which can be solidified by irradiating it with a high-energy beam. In particular, the invention relates to a powder pre-heating process using an electron beam.

BACKGROUND ART

Equipment for producing a three-dimensional object layer by layer using a powdery material which can be solidified, or fused together, by irradiating it with a high-energy beam of electromagnetic radiation or electrons are known from e.g. US4863538, US5647931 and SE524467. Such equipment include for instance a supply of powder, means for applying a layer of powder on a vertically adjustable platform or working area, and means for directing the beam over the working area. The powder sinters or melts and solidifies as the beam moves over the working area.

When melting or sintering a powder using a high-energy beam, it is important to avoid exceeding the vaporization temperature of the powder, since otherwize the powder will merely vaporize instead of forming the intended product. US2005/0186538 discloses a method focusing on this problem. In this method a laser beam repeatedly is directed to the same powder target area during the melting/sintering phase as to stepwize raise the powder temperature. This way, a too high powder temperature is avoided.

When using an electron beam instead of a laser beam, the situation is in some ways different. As the electron beam hits the powder, a charge distribution develops around the electron target area. If the charge distribution density exceeds a critical limit, an electrical discharge will occur since the powder particles will repel each other. A result of such a discharge is that the structure of the powder layer will be destroyed. Applying the method according to US2005/0186538 to a powder melting/sintering device equipped with an electron beam is likely to give a poor result since no measures are taken in that method to avoid such discharges.

One solution to the problem of avoiding discharges is to add conductive material, such as carbon, to the powder as to increase the electrical conductivity of the powder. Disadvantages of this solution are, however, that the solidifying process of such a powder mixture may be difficult to control and that the properties of the formed product may be affected in a negative way. For instance, the mechanical strength may be decreased.

DISCLOSURE OF INVENTION

The object of this invention is to provide a method and device for layer-wize production of three-dimensional objects from a powdery material, which method and device allow for a controlled and proper fusing together of the powdery material, and which are well suited for both an electron beam and a laser beam. This object is achieved by the method and device defined in independent claims 1 and 11 , respectively. The dependent claims contain advantageous embodiments, further developments and variants of the invention.

The invention concerns a method for producing three-dimensional objects layer by layer using a powdery material, which can be solidified by irradiating it with a high-energy beam. The invention is characterized in that the method comprises a pre-heating step with the general purpose of pre-heating the powdery material in a homogeneous manner, followed by a solidifying step with the general purpose of fusing together the powdery material, wherein the pre-heating step comprises the sub-step of scanning a pre-heating powder layer area by scanning the beam along paths distributed over the pre-heating powder layer area, wherein consecutively scanned paths are separated by, at least, a minimum security distance, said minimum security distance being adapted to prevent undesirable summation effects in the pre-heating powder layer area from said consecutively scanned paths.

An advantage of the inventive method is that the pre-heating step allows the powder layer to be homogenously heated up so as to avoid having too large temperature gradients in the interface between melted metal and powder in the subsequent solidifying step. By using a security distance adapted to prevent summation effects from consecutively scanned paths, it is possible to avoid that the energy deposited to the powder during scan of a first path adds to the energy deposited during scan of a second path scanned directly after the first path. Thus, large temperature gradients can be avoided also during the pre-heating stage.

When using an electron beam the pre-heating has a further advantageous effect in that it increases the electrical conductivity of the powder. This has in turn the effect that a high beam current can be used in the subsequent solidifying step. Also the security distance has an additional advantage when using an electron beam in that it eliminates the risk of forming a too large charge density in the relatively cool powder during the pre-heating step. Thus, powder discharge is prevented.

The inventive use of the beam for pre-heating the powder has several advantageous compared to the rather obvious alternative to heat up an entire powder bed using e.g. heating elements. One advantage is that no further heating equipment is needed. Another advantage is that only the part of the powder bed that really needs to be heated, i.e. a fraction of the upper layer of the powder bed, is actually heated. This makes the process very efficient.

In a first advantageous embodiment of the inventive method the pre-heating step further comprises the sub-step of re-scanning the pre-heating powder layer area. This way the pre-heating area can be gradually and homogeneously heated up. Preferably, the paths followed during a re-scan of the pre-heating powder layer area are displaced an interspacing distance in relation to the paths followed during a previous scan of the pre-heating powder layer area, wherein the interspacing distance is less than the minimum security distance. This way it is possible to obtain a homogeneously pre-heated powder layer area also in situations where it is necessary to use a scanning pattern where the paths are physically separated by a distance that is shorter than the minimum security distance, i.e. in situations where additional, more closely positioned paths are needed.

In a second advantageous embodiment of the inventive method the power of the beam is increased during the pre-heating step. This has the advantage that the beam power initially can be kept at a sufficiently low level to avoid large charge densities and/or temperature gradients, but also that the beam power increases as the temperature of the powder increases as to speed up the pre-heating process as much as possible. In a preferred variant of the inventive method, the power of the beam is increased stepwize between consecutive scans or re-scans of the pre-heating powder layer area. This makes it relatively easy to control the process and allows the powder preheating area to be heated in a uniform manner.

In a third advantageous embodiment of the inventive method the beam is an electron beam wherein the beam power is increased by increasing the beam current.

Preferably, the paths form substantially straight and parallel lines. Such paths simplifies the labour of finding a path pattern and path scanning order that works in practice, considering summation effects etc. Using straight and parallel paths also simplifies the control of the beam during scanning. To further simplify the process, the paths are preferably scanned from one end to the other. In a fourth advantageous embodiment of the inventive method the preheating powder layer area is larger than, and thereby forms a security margin with respect to, a corresponding part of the powder layer that is to be fused together in the subsequent solidifying step. Such a security margin ensures that the whole product area, including its close surroundings, is properly preheated, i.e. such that the temperature and electrical conductivity of the powder layer do not change abruptly at the outer border of parts that are to be fused together. Without such a security margin it is likely that problems will arise due to a too large temperature gradient and/or a too large charge density.

The invention also concerns a device adapted to be operated according to the inventive method.

BRIEF DESCRIPTION OF DRAWINGS

In the description of the invention given below reference is made to the following figure, in which:

Figure 1 shows, in a schematic view, an example of a known device for producing a three-dimensional product to which device the inventive method can be applied.

Figure 2 shows an example of a first preferred embodiment of the inventive method, and

Figure 3 shows examples on how to apply the inventive method to different product shapes.

EMBODIMENT(S) OF THE INVENTION

Figure 1 shows an example of a known device 1 for producing a three- dimensional product. The device 1 comprises a vertically adjustable work table 2 on which a three-dimensional product 3 is to be built up, one or more powder dispensers 4, means 28 arranged to distribute a thin layer of powder on the work table 2 for forming a powder bed 5, a radiation gun 6 in the form of an electron gun for delivering energy to the powder bed 5 as to fuse together parts of the powder bed 5, deflection coils 7 for guiding the electron beam emitted by the radiation gun 6 over said work table 2, and a control unit 8 arranged to control the various parts of the device Un a typical work cycle, the work table 2 is lowered, a new layer of powder is applied onto the powder bed 5, and the electron beam is scanned over selected parts of the upper layer 5' of the powder bed 5. In principal, this cycle is repeated until the product is finished. An expert in the field is familiar with the general function and composition of devices for producing a three-dimensional product, both with regard to the type outlined in figure 1 and to devices equipped with a laser gun instead of an electron gun.

In the case where an electron beam is used, it is necessary to consider the charge distribution that is created in the powder as the electrons hit the powder bed 5. The invention is, at least partly, based on the realization that the charge distribution density depends on the following parameters: beam current, electron velocity (which is given by the accelerating voltage), beam scanning velocity, powder material and electrical conductivity of the powder, i.e. mainly the electrical conductivity between the powder grains. The latter is in turn a function of several parameters, such as temperature, degree of sintering and powder grain size/size distribution.

Thus, for a given powder, i.e. a powder of a certain material with a certain grain size distribution, and a given accelerating voltage, it is possible, by varying the beam current (and thus the beam power) and the beam scanning velocity, to affect the charge distribution.

By varying these parameters in a controlled way, the electrical conductivity of the powder can gradually be increased by increasing the temperature of the powder. A powder that has a high temperature obtains a considerably higher conductivity which results in a lower density of the charge distribution since the charges quickly can diffuse over a large region. This effect is enhanced if the powder is allowed to be slightly sintered during the pre-heating process. When the conductivity has become sufficiently high, the powder can be fused together, i.e. melted or fully sintered, with arbitrary values of the beam current and beam scanning velocity.

A preferred embodiment of the inventive method, wherein the conductivity is increased without creating discharges, is shown in figure 2. Here, the beam is scanned along paths distributed in a certain pattern over a part of the powder bed 5 that is to be solidified, for the purpose of pre-heating the powder. The part of the upper layer 5' of the powder bed 5 subjected to preheating is denoted the pre-heating powder layer area 10, or only pre-heating area 10 (see also figure 3). Reference numbers L_x and L_y denotes the sides of the, in this example rectangular, pre-heating area 10. The beam follows the paths, indicated by straight and parallel lines P1.1 , P1.2 etc, from the left to the right, i.e. from x = 0 to x = L_x. On the left side of the lines/paths, a code of each path is given. On the right side of the lines/paths, the order in which the paths are scanned is given. Thus, the first path to be scanned is P1.1 , the next path is P2.1 , after that path P3.1 and so on. Such consecutively scanned paths are physically separated by a security distance ΔY that will be further discussed below.

Depending on the particular conditions, such as dimensions of pre-heating area 10, beam power and beam scanning velocity, it may be necessary to use a scanning pattern where the paths are physically separated by a distance that is shorter than the minimum security distance ΔY in order to obtain a homogeneously pre-heated powder layer area 10. Figure 2 shows an example of such a case where additional, more closely positioned paths are needed. The additional paths are indicated with codes P1.2, P1.3 etc. Adjacent paths, such as P1.20 and P2.1 or P3.2 and P3.3, are physically separated by an interspacing distance δY. As can be seen on the right side of the lines/paths in figure 2, adjacent paths are not scanned in a consecutive order in order to still separate consecutively scanned paths by the security distance ΔY.

In the example shown in figure 2, the pre-heating area 10 can be seen as divided into five sub-areas, P1 to P5, and in each sub-area the number of paths to be scanned are 20; e.g. P1.1 to P1.20 in the first sub-area P1. More generally, the paths can be denoted PM. N, wherein M is the number of the sub-area and N is the number of a particular path in the sub-area M. In figure

2, M goes from 1 to 5 and N goes from 1 to 20, which leads to a total number of 100 paths to be scanned. The values of M and N may be varied depending on e.g. the size of the pre-heating area 10 and on the desired pre-heating temperature and/or desired degree of pre-sintering of the powder.

As given by the path scanning order in figure 2, the pre-heating area 10 is scanned several times in the y-direction, in this example 20 times. The first time the pre-heating area 10 is scanned, the scanning procedure is such that the first path PM.1 in each sub-area P1 to P5 is scanned. When this initial step is finished, the pre-heating area 10 is re-scanned by scanning the second path PM.2 in each sub-area P1 to P5. In the next re-scan, the third path PM.3 in each sub-area P1 to P5 is scanned, and so on. This procedure may be seen as one single scanning pattern, comprising the N:th path of each sub-area P1-P5, which single scanning pattern is displaced in the y- direction, i.e. downwards in figure 2, a distance corresponding to the interspacing distance δY once the scan of all paths in the single scanning pattern has been completed. In other words, the paths of a re-scan are displaced in parallel a distance δY relative to the paths of the previous scan. The reference N denotes the order of the scan or re-scan of the pre-heating area 10, wherein N starts with 1 (for the first scan) and goes to a maximum value that, in this example, is 20 (for the last scan). Below, this maximum value of N is denoted N_r.

Each scanning or re-scanning of the pre-heating area 10 has the effect of increasing the temperature of the powder bed that in turn has the effect of increasing the electrical conductivity of the powder. The beam current can therefore be increased after each (re-)scanning procedure. How much the beam current can be increased between the scans depends on how much the conductivity can be increased in the preceding scan.

It is important that the beam current, the beam scanning velocity and the paths to be scanned are adapted such that the charge density around the position where the beam hits the powder is prevented from exceeding a critical limit above which discharge will occur.

A general function for describing the charge density that develops in the powder in an arbitrary scanning procedure will be a rather complex function of time and beam position since the charge density generated along one scanned path will be affected by the charge density generated along another scanned path if these paths are not very well separated in space and time. Thus, summation effects between different paths must be taken into account.

In a predetermined scanning procedure using straight and parallel scanning paths, similar to those shown in figures 2 and 3, summation effects are much easier to control. For a single, straight path the charge density depends the fraction I/V_s, where I is the beam current and V₃ is the beam scanning velocity relative to the powder bed. If this fraction is too high, too much charge will be deposited to the powder per path length unit. From a production point of view it is desirable to increase the temperature in an efficient manner to minimize the time required for pre-heating the powder. Thus, the beam current and the beam velocity should be as high as possible without exciding the critical limit in charge density. However, since charges will remain around a scanned path for some time the summation of charge density between different scans has to be considered. It is important that the beam does not return to the same position, or to the close vicinity of the same position, until a certain minimum time period to has elapsed. Thus, for a given path length in a pre-heating area 10 the beam scanning velocity is not solely determined by the fraction I/V_s but also by the time period t_othat has to elapse before the beam can return to the same position. As the charge density decreases not only with time but also with distance from the previously scanned position, the required time period that must be allowed to elapse before a certain position of the powder layer can be scanned decreases with increasing distance from the previously scanned position. In a first order approximation this distance-dependent security time period, t_p , can be considered to be independent of beam current and set to:

t_p = to- k_r*r,

where t₀ is the time that has to elapse before the beam can return to the same position as it was at t = 0, r is the distance between the beam position at t = 0 and the new position at time t, and k_r is a proportional factor. Here, t_p is assumed to have values between 0 and to, which means that summation effects are considered to be negligible for sufficient large values of r.

Consequently, the scanning of the paths of the pre-heating area 10 must be arranged such that the paths become sufficiently separated in time and/or space so as to avoid undesirable charge summation effects in the preheating powder layer area 10. This holds both for charge summation effects as discussed above and for energy summation effects wherein the amounts of energy deposited along two paths add together so as to locally raise the temperature too much. Consecutively scanned paths, such as P4.2 and P5.2 in figure 2, must be physically separated to a higher extent than other scanned paths since the time period elapsed between consecutively scanned paths is shorter (provided that the paths are of equal length and an equal time period between start of the scan of each individual path).

From a given beam scanning velocity, V₅, and a given length of the paths, L_x, it is possible to convert the required distance-dependent security time period tp to a minimum security distance ΔY, which is easier to handle in practice than a minimum time period. The required length of this distance ΔY depends on how fast the beam returns to x = 0. Thus, ΔY increases with decreasing length of the paths L_x and with increasing beam scanning velocity V₃. In figure 2, consecutively scanned path, such as P4.2 and P5.2, are separated by the minimum security distance ΔY.

As described above, a certain time period must be allowed to elapse before different paths can be scanned. In order to reduce the total time required for pre-heating the powder, it is important that the beam scans parts of the pre- heating area 10 that are not subject to a "period of rest" set by t_p.

In the example below the following parameters are used:

L_x, Ly = lengths of the sides of the pre-heating powder layer area 10,

V₃ = beam scanning velocity, I₀ = initial beam current,

Δl = beam current increase between re-scanning of the pre-heating area 10, Np = number of times the pre-heating area 10 is scanned, ΔY = distance between two consecutively scanned paths; minimum security distance, and δY = distance between two adjacent path; interspacing distance.

For a given powder layer area, i.e. where values of L_x, L_y are given, it is possible to empirically obtain the values for V_s, lo, Δl, N₁-, ΔY and δY that are required for a proper pre-heating of the powder area in question.

Table 1 shows an example of proper values of V₈, lo, Δl, N₁-, ΔY and δY for a certain powder layer area (L_x, L_y), a certain accelerating voltage (60 kV), and a certain, commercially available, powder (gas atomized ELI TΪ6AI4V). Table 1

Assuming that the time required for the beam to "jump" between different paths is negligible (which normally is a fair assumption since the "jump" velocity of a beam normally is much greater than its scanning velocity) and assuming that linear relationships are valid, it is possible to use the parameter values in table 1 for producing relationships that can be used for arbitrary values of L_x and L_y. To obtain the most accurate empirical values, these values should be produced using a powder area that is as small as possible, i.e. the values of L_x and L_y should be as small as possible. However, approximate empirical values, that might be sufficiently accurate, can be obtained in a faster way by starting out with a larger powder area (larger L_x and L_y). Preferably, the same value of δY is used independently of the values of L_x and L_y because δY also affects the surface finish of the completed three-dimensional product. It is also important that the total energy deposited per area unit is evenly distributed in order to keep the temperature as even as possible over the powder area in question irrespective of the values of L_x and L_y.

With reference to table 1 , the following relationships and limiting parameters are valid: to = (L_y/ΔY) -L_x/V_s, kr = (to - L_x/V_s )/ΔY, ki = lo/Vs, k₂ = Δl/ V_s, and k₃ = (I₀ + Nr-ΔI) 'N_r/( V_s -δY^« 2),

where to is the minimum time period that must be allowed to pass before the beam returns to (the close vicinity of) a previously scanned path (i.e. to is the time period that has to elapse before the beam can return from e.g. line PM. N to PM.N+1 ); k_r is the factor used to determine the time period that must be allowed to pass before the beam returns to x = 0 at a distance ΔY from a previously scanned path; ki is proportional to the maximum amount of charge deposited per mm of the paths during the first scan of the pre-heating area 10; k₂ is proportional to the maximum charge deposition increase per mm for each re-scan of the pre-heating area 10; and k₃ is proportional to an average energy deposition per mm² required for keeping the powder surface at a certain temperature.

Here, to and k_r are minimum values, whereas ki and k₂ are maximum values that should not be exceeded. The factor k₃ is a form of guideline value but can be seen as a maximum value that should not be exceeded for the purpose of speeding up the process.

Values of these limiting parameters can be obtained by using the empirically obtained values in table 1. After having obtained these limiting parameters, they can be used to calculate the five unknown parameters V₅, lo, Δl, N_r and ΔY for arbitrary values of L_x and L_y, as long as δY is kept at almost the same value. Some care has to be taken since the fractions L_y/ΔY and ΔY/δY have to be integers. Thus, the parameters may be determined in an iterative way where, for instance, L_x is kept fixed whereas L_y and δY are allowed to vary somewhat.

The task of obtaining empirical values, such as those given in table 1 , for other types of powders on the basis of the information given in this text, can be considered to be routine work for a man skilled in the art. A general rule is that t_Ol and thus t_p, increase with decreasing conductivity of the powder. Thus, for a powder with a low conductivity, large values for L_x, L_y, V_s, N₃ and ΔY might be necessary; in conjunction with low values for I₀ and Δl.

As described above, pre-heating of the powder layer may be performed over a rectangular powder layer area that encloses all parts of the powder that are to be fused together. This may, however, be an inefficient approach since, depending on the form of the product to be produced; an unnecessarily large powder area might be heated up.

Figure 3 schematically shows, in a vertical view, three examples of different shapes of powder layers that are to be fused together as to form a part of the product 3. Figure 3 also shows the corresponding pre-heating areas 10 (dashed lines) and some selected paths P (thin solid lines) to be followed during the pre-heating scan. Figure 3a shows a product 3 that, at least in this particular layer, has an oval shape with a hole in the middle, whereas figures 3b and 3c show products 3 having a rectangular and a circular shape, respectively.

As can be seen in figure 3, the shapes of the pre-heating areas 10 have the same principal shape as the product, i.e. as the shape of the powder layer to be fused, but the pre-heating areas 10 are enlarged as to enclose the parts that are to be melted together. The size of each pre-heating area 10 is adapted such that a certain security margin 12 is formed with respect to the corresponding part 3 of the powder layer that is to be fused together. The security margin 12 should be sufficient for ensuring that the whole product area 3, including its close surroundings, is properly pre-heated, i.e. such that the temperature and electrical conductivity of the powder layer do not change abruptly at the outer border of the parts that are to be fused together. For the Ti6AI4V-powder mentioned in connection with table 1 , the security margin 12 should be at least 6 mm. Generally, the magnitude of the security margin 12 should be increased with decreasing heat and/or electrical conductivity of the powder.

As can be seen in figures 3a and 3c, the lengths of the paths may vary. In such cases it may be needed to adjust some parameters, such as the minimum security distance ΔY, to take into account that some paths takes shorter time to scan.

If the powder used has a very low electrical conductivity, and/or if the distance is unusually long between the parts of the powder layer that are to be fused together, it may be necessary to pre-heat also parts of the powder layer that are not to be fused together in order to allow the beam to "jump" between the parts that are to be fused together. Otherwise, when repeatedly jumping over the same not-to-be-fused area, the charge distribution density in this area may exceed the critical value.

The term "jump" refers to the situation when the beam quickly is moved from one position to another, for instance from an end position of a path to a starting position of the next path to be scanned. In some applications it may be favourable to "jump" instead of switching the beam on and off.

Once the pre-heating method step has been concluded, a solidifying method step may follow in which the beam energy can be further increased as to melt or sinter the powder grains together. By executing the pre-heating method step in a controlled and accurate manner, it is possible to ensure that the subsequent solidifying step will be properly carried out.

Although most of the advantages of the inventive method can be achieved when using an electron beam, the method is beneficial also in laser beam applications. One example is that the inventive method is capable of producing a homogeneously sintered powder layer area. Such sintered areas will increase the heat conductivity in the powder and thus minimize the possibility of having too large temperature gradients in the interface between melted metal and powder in the subsequent melting step.

The invention is not limited by the embodiments described above but can be modified in various ways within the scope of the claims. For instance, if the ratio L_x/V_s is large, it is possible to position the subsequent path close to the path just scanned. In such a case the security distance ΔY may be set equal to the interspacing distance δY, i.e. the same paths are scanned in each re- scan.

It is further possible to scan the paths in a different order than what is described in relation to figure 2. For instance, the first path in each sub-group (P1.1 , P2.1 , etc.) could be scanned several times before the second path in each sub-group (P1.2, P2.2, etc.) is scanned. On some occasions, in particular if the fraction L_x/V_s is large, it may also be possible to scan the same path several times without scanning any other paths in between.

Moreover, the paths do not necessarily have to be straight and parallel lines. However, such a path pattern simplifies the labour of finding a path pattern and path scanning order that works in practice, considering summation effects etc. Using straight and parallel paths also simplifies the control of the beam during scanning.

Claims

1. Method for producing three-dimensional objects (3) layer by layer using a powdery material (5) which can be solidified by irradiating it with a high- energy beam, characterized in that the method comprises a pre-heating step with the general purpose of pre-heating the powdery material (5) in a homogeneous manner, followed by a solidifying step with the general purpose of fusing together the powdery material, wherein the pre-heating step comprises the sub-step of: - scanning a pre-heating powder layer area (10) by scanning the beam along paths (P1.1 - P5.20) distributed over the pre-heating powder layer area (10), wherein consecutively scanned paths (PM. N, P(M+1).N) are separated by, at least, a minimum security distance (ΔY), said minimum security distance (ΔY) being adapted to prevent undesirable summation effects in the pre- heating powder layer area (10) from said consecutively scanned paths.

2. Method according to claim 1 , characterized in that the pre-heating step further comprises the sub-step of: - re-scanning the pre-heating powder layer area (10).

3. Method according to claim 2, characterized in that the paths followed during a re-scan of the pre-heating powder layer area (10) are displaced an interspacing distance (δY) in relation to the paths followed during a previous scan of the pre-heating powder layer area (10), wherein the interspacing distance (δY) is less than the minimum security distance (ΔY).

4. Method according to anyone of the above claims, characterized in that the power of the beam is increased during the pre-heating step.

5. Method according to claim 4 and claim 2 or 3, characterized in that the power of the beam is stepwize increased between consecutive scans or re-scans of the pre-heating powder layer area (10).

6. Method according to claim 4 or 5, characterized in that the beam is an electron beam and that the beam power is increased by increasing the beam current.

7. Method according to anyone of the above claims, characterized in that the paths (PM. N) are scanned from one end (x = 0) to the other (x = L_x).

8. Method according to anyone of the above claims, characterized in that the paths (PM. N) are substantially parallel.

9. Method according to anyone of the above claims, characterized in that the paths (PM. N) form substantially straight lines.

10. Method according to anyone of the above claims, characterized in that the pre-heating powder layer area (10) is larger than, and thereby forms a security margin (12) with respect to, a corresponding part (3) of the powder layer that is to be fused together in the subsequent solidifying step.

11. Device (1) for producing three-dimensional objects (3) layer by layer using a powdery material (5) which can be solidified by irradiating it with a high-energy beam, characterized in that it is adapted to be operated by a method according to anyone of the above claims.