WO2023152015A1

WO2023152015A1 - Method for laser welding a bipolar plate for a fuel cell, with power-density distribution varying cyclically over time in the region of the molten pool

Info

Publication number: WO2023152015A1
Application number: PCT/EP2023/052521
Authority: WO
Inventors: Oliver BOCKSROCKER; Nicolai Speker; Tim Hesse; Philipp Scheible
Original assignee: Trumpf Laser- Und Systemtechnik Gmbh
Priority date: 2022-02-10
Filing date: 2023-02-02
Publication date: 2023-08-17
Also published as: DE102022103167A1

Abstract

A method for laser welding a bipolar plate (1) for a fuel cell, wherein two plate parts (1a, 1b) are welded to one another along at least one weld seam (2, 2a, 2b, 2c), wherein the laser welding is performed with a laser beam (5), wherein the laser beam (5) has, in a plane (E) of a surface (4) of the two plate parts (1a, 1b), a basic movement component (GBK) with a rate of advancement VS in a welding direction (SR) along a welding curve (8) in relation to the plate parts (1a, 1b), and the welding curve (8) runs along the weld seam (2, 2a, 2b, 2c), wherein the laser beam (5) produces a molten pool (7) in the plate parts (1a, 1b), and the laser beam (5) brings about a power-density distribution (LDV) of laser radiation in the plane (E) of the surface (4) of the two plate parts (1a, 1b) in the region of the molten pool (7), and wherein the laser beam (5) comprises one or more part-beams (11a, 11b), is characterized in that the power-density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (1a, 1b) is varied cyclically over time in the region of the molten pool (7). The invention provides a method by which low-defect weld seams with a high level of fluid impermeability can be produced on a bipolar plate at a high welding speed.

Description

Process for laser welding a bipolar plate for a fuel cell, with power density distribution that varies cyclically over time in the area of the weld pool

The invention relates to a method for laser welding a bipolar plate for a fuel cell, two plate parts being welded to one another along at least one weld seam, the laser welding being carried out with a laser beam, the laser beam having a basic movement component with a feed rate in a plane of a surface of the two plate parts VS in a welding direction SR. along a welding curve relative to the plate parts, and the welding curve runs along the weld seam, the laser beam generating a molten pool in the plate parts, and the laser beam in the plane of the surface of the two plate parts in the region of the molten pool effects a power density distribution of laser radiation, and wherein the laser beam comprises one or more sub-beams. Such a method has become known from the subsequently published German patent application DE 10 2021 113 834.5.

In fuel cells with several cells layered to form a stack, bipolar plates are used for the distribution of gases, in particular hydrogen and oxygen, the removal of water (water of reaction), the gas-tight separation between adjacent cells and the seal to the outside and cooling. In addition, the bipolar plate on the hydrogen side absorbs the electrons that are released and feeds them back to the oxygen side.

Such bipolar plates can have two metallic plate parts that are welded together. On the one hand, weld seams must be fluid-tight in order to direct gases and water in defined paths. On the other hand, weld seams serve to connect the two plate parts electrically and mechanically.

It has become known from the subsequently published German patent application 10 2021 113 834.5 to produce at least one peripherally closed first weld seam with a first seam width and at least one second weld seam with a second seam width when laser welding plate parts of a bipolar plate, the second seam width being larger than the first seam width.

In order to use bipolar plates economically, it is desirable to weld the plate parts of the bipolar plates together in the shortest possible time. This requires high feed speeds of the laser beams with which the laser welding takes place. At high feed speeds, however, melt accumulations (also known as humps) can form in the wake of the weld seam ("humping"). The melt accumulations can lead to faulty connection of the plate parts and a leak in the weld seam. As a result, the humping limits the maximum achievable feed rate (welding speed). DE 10 2016 204 578 B3 proposes using power modulation on the laser beam during laser beam welding of a steel workpiece to avoid hot cracks. A modulation frequency f of the power modulation is selected in such a way that for a normalized, characteristic oscillation frequency Aco of the melt pool and a normalized modulation frequency A, the following applies:

A > 2.2*Aco, and A < 8.5*Aco, with ^v , with v: feed rate of the laser beam relative to the workpiece, and d diameter of a laser beam

focal spot of the laser beam, and ^Vc ° determined from a test measurement with the laser beam without modulation of the laser power, with fco ^test : measured characteristic oscillation frequency in the test measurement; df,co ^test : diameter of the laser beam focal spot in the test measurement; v _co ^test : Feed speed of the laser beam relative to the workpiece during the test measurement.

From A. Heider et al., "Power modulation to stabilize laser welding of copper", Journal of Laser Applications 27, 022003 (2015); doi : 10.2351/1.4906127, it has become known to stabilize the welding process when welding copper, by modulating the laser power of the laser used for welding.The influence of the average power, the modulation amplitude, the welding speed, the focus diameter and the modulation frequency on the welding quality is examined in more detail.

In A. Jahn et al., "High dynamic beam shaping by piezo driven modules for efficient and high quality laser beam cutting and welding", conference paper "Lasers in Manufacturing Conference 2019", Wissenschaftliches Gesellschaft Lasertechnik e. V. (WLT), a concept for beam shaping optics of laser beams for laser welding and laser cutting processes is described, in which piezo actuators are used for fast focus modulation of the laser beam in the z-direction with working frequencies above 2.5 kHz. Furthermore, first experimental results are presented that show the influence of the modulation in the z-direction on the process behavior for typical laser welding and laser cutting processes. US 2004/00262381 A1 proposes to produce a high-quality weld seam when laser welding steel plates, when carrying out the laser welding with a laser beam whose power is pulse-modulated, the power with a frequency which corresponds to a natural frequency of a molten pool generated by the laser beam , to vary periodically.

It is the object of the present invention to provide a method with which low-defect weld seams with high fluid tightness can be produced in a bipolar plate at a high welding speed.

This object is achieved according to the invention by a method of the type mentioned at the outset, which is characterized in that the power density distribution in the plane of the surface of the two plate parts in the region of the melt pool is varied cyclically over time.

If, during laser welding with a laser beam, a high feed rate VS in the welding direction SR. Applied to the laser beam, (if no countermeasures are taken) melt accumulations can occur in the wake of the weld seam behind the weld pool. This process of formation of lumps of melt is also referred to as humping, and the lumps of melt are also referred to as humps. The humping is caused by a movement of the melt pool generated during laser welding against the welding direction SR and rapid solidification of the melt. The melt accumulations occur periodically in the solidified weld.

The present invention proposes a temporally cyclical variation of the power density distribution in the plane of the surface (facing the laser beam) of the two plate parts in the region of the melt pool. By varying the power density distribution in the area of the melt pool, the Flow of the melt in the melt pool can be changed counter to the feed direction VS, and in particular the formation of wave crests and valleys can be reduced. This can reduce or even eliminate humping. The melt pool dynamics can be calmed down.

Within the scope of the invention, welding can be carried out with a high feed rate VS and a solidified weld seam can be produced with only weakly pronounced or even completely no melt accumulations. The production speed for the bipolar plates can therefore be increased in comparison to laser welding with a power density distribution that is constant over time in the plane of the surface of the two plate parts in the area of the melt pool. Furthermore, by suppressing the accumulations of melt, reliably fluid-tight weld seams can be produced. In addition, weld spatter (also referred to as splatter for short) and ejections during welding are reduced.

The cyclical variation of the power density distribution in the plane of the surface of the two plate parts in the area of the molten pool can take place in particular in that at least one partial beam of the laser beam undergoes a temporally cyclical power change and/or that at least one partial beam of the laser beam changes the basic movement component a time-cyclical additional movement component ("pendulum movement component") is superimposed. It should be noted that a pendulum movement component typically has a much higher mean speed in terms of absolute value than the basic movement component, usually by a factor of at least 10 or a factor of 50.

The (cyclically changing) power density distribution of the laser radiation is observed in the area of the molten pool (i.e. relative to the molten pool), with the molten pool progressing essentially with the basic movement component along the welding curve. In other words, the frame of reference for observing the (cyclically changing) power density distribution moves with the basic motion component. The power density distribution is observed in the plane of the workpiece surface (plane of the upper side of the upper plate part facing the laser), i.e. essentially on the surface of the melt pool, and in the area of the melt pool, i.e. within the extent of the surface of the melt pool.

It should be noted that the time-cyclic variation of the power density distribution in the plane of the surface of the plate parts in the area of the melt pool is basically accompanied by a variation of a power density distribution in the volume of the melt pool (and in particular over the depth of the melt pool), which can also be observed if desired. However, the variation of the power density distribution in the volume of the melt pool is not discussed in detail below.

The temporally cyclical change in the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool preferably includes at least one portion that changes over time in the direction and counter to the direction of the basic movement component.

The metallic plate parts of the bipolar plate are arranged to overlap for welding (typically with a congruent, aligned edge contour). The plate parts of the bipolar plates typically have a profiling, through which channels for cooling water and, on the outside, ducts for reaction water and/or gases, in particular hydrogen and oxygen, are formed between the plate parts. In addition, the bipolar plates typically have one or more openings with which gas can be transported in the stack direction in the fuel cell stack. Within the scope of the invention, closed welds are typically made at the outer edge of the plate parts and around all openings, and non-closed welds are also made distributed over the surface of the bipolar plates. The fluid tightness, in particular the gas tightness, is particularly important for the closed weld seams, and the mechanical and electrical connection is particularly important for the open weld seams. The invention can can be used for both closed welds and non-closed welds.

In general, the fluid mechanics in the area of the melt pool can be specifically influenced via the inventive cyclical variation of the power density distribution in the plane of the surface of the two plate parts, and the local welding depth can also be influenced if necessary, and humping can be suppressed.

variation in performance

In a particularly preferred variant of the method according to the invention, it is provided that a laser power of at least one partial beam of the laser beam is varied cyclically over time in order to vary the power density distribution in the plane of the surface of the two plate parts in the region of the melt pool.

As a result, the movement of the melt in the melt pool can be influenced in a simple manner and humping can be counteracted. The instantaneous size or depth of a vapor capillary can be influenced by varying the power density. Depending on the requirements during welding, the entire laser beam, or only a partial beam of the laser beam, or several partial beams of the laser beam can be cyclically varied over time. The modulation depth MT=1-(Pmin/Pmax) is typically between 5% and 95%, with Pmin: minimum laser power of the at least one partial beam during a cycle and Pmax: maximum laser power of the at least one partial beam during a cycle. A temporally cyclical variation of the laser power of the at least one partial beam can be combined with a spatial oscillation of the at least one partial beam (or its partial laser spots) (e.g. with a transverse movement component, see below, in particular in the area of the reversal points a reduced laser power is applied to the transverse movement component in order to avoid a molar tooth profile of the weld seam). A further development of this variant provides that the laser power of the at least one partial beam is varied sinusoidally or rectangularly or triangularly or as a combination thereof. All of these variation options are suitable for optimizing the melt flow in the melt pool, which enables fast processing of the plate parts.

A further development of this variant is also preferred, in which the laser power of the at least one partial beam is varied sinusoidally in areas of power troughs and triangularly in areas of power peaks, and a power jump is set up at transitions from sinusoidal and triangular variations. Through this combination of the different modes of vibration, the melt flow in the melt pool can be optimized particularly well and the plate parts can be processed quickly.

A further development is also preferred in which a laser power of the laser beam is varied cyclically overall over time in order to vary the power density distribution in the plane of the surface of the two plate parts in the region of the melt pool. The time-cyclical variation of the laser power of the laser beam as a whole is technically particularly easy to implement. If necessary, an instantaneous size or depth of a vapor capillary can be influenced by varying the laser power.

Also preferred is a further development characterized in that, in order to vary the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool, a power distribution between at least two partial beams is varied cyclically over time, with the total power of the laser beam being kept constant over time. A particularly uniform weld seam can be welded with the same overall performance. At the same time, due to the cyclical variation of the power distribution between the at least two partial jets, the melt flow in the melt bathroom can be specifically influenced and optimized. The at least two partial beams can, for example, be partial beams lying next to one another or partial beams lying one inside the other.

Variation of place (Wöbbel n)

In a further preferred variant, it is provided that, in order to vary the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool, a transverse movement component is superimposed on at least one partial beam in the plane of the surface of the plate parts, so that at least one partial laser spot of the at least one partial beam is moved back and forth cyclically in time transversely to the welding curve. The movement in the melt pool can be influenced by the temporally cyclic transverse movement component of the at least one partial beam, in particular in order to suppress the formation of melt accumulations. In addition, the seam width or the welding depth can also be influenced if necessary. Depending on the requirements during welding, the entire laser beam, or only a partial beam of the laser beam, or several partial beams of the laser beam can be cyclically moved back and forth in time.

A further development of this variant provides that, in order to vary the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool, a transverse movement component is superimposed on the entire laser beam, so that a laser spot of the laser beam moves back and forth cyclically transverse to the welding curve becomes. The time-cyclic back and forth movement of the entire laser beam is technically easy to implement.

A variant is particularly preferred in which, in order to vary the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool, a longitudinal movement component is superimposed on at least one partial beam in the plane of the surface of the plate parts, so that at least one partial laser spot of the at least a partial beam parallel to the welding curve borrowed is moved back and forth cyclically. The movement in the melt pool can be influenced particularly efficiently by the chronologically cyclic longitudinal movement component of the at least one partial beam, in particular in order to suppress the formation of melt accumulations. In addition, the melt pool length or the welding depth can also be influenced if necessary. Depending on the welding requirements, the entire laser beam, only a partial beam of the laser beam, or several partial beams of the laser beam can be cyclically moved back and forth in time.

In a further development of this variant, it is provided that, in order to vary the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool, a longitudinal movement component is superimposed on the entire laser beam, so that a laser spot of the laser beam parallel to the welding curve cyclically backwards in time and is moved. The time-cyclic back and forth movement of the entire laser beam is technically easy to implement.

A further development is also preferred in which, for a mean forward movement speed v in front of the longitudinal movement component in the welding direction SR. and for an average reverse movement speed vrück of the longitudinal movement component against the welding direction SR, the following applies: | vback| > |vvor| . In this way, the melt can be decelerated against the welding direction SR or accelerated in the welding direction SR. With this procedure, the melt flow in the melt pool can be calmed down particularly effectively, and the formation of melt accumulations can be well suppressed. |vrück| is often still valid >l,5*|vvor| or also |vrück| >2*|vvor| or also 0.2*|vback| < |vvor| <0.8*|vback|, preferably |vforward| =0.5*| vback| . The relative speed between the at least one partial beam and the bipolar plate on the workpiece surface can be reduced as a result of the mean return movement speed, which is higher in terms of absolute value, in particular during the forward movement of the longitudinal movement components. It should be noted that |vback| < |vvor| or |vback| = |vbefore| can be chosen. Also preferred is a further development which is characterized in that the at least one partial beam, on which the longitudinal movement component is superimposed, moves in the welding direction SR. has an average forward movement laser power Pvor and, in the case of movement counter to the welding direction SR, an average backward movement laser power Prück, where: Pvor<Prück. In this way, the melt can also be decelerated against the welding direction SR or accelerated in the welding direction SR. This procedure can also very effectively calm the melt flow in the melt pool and suppress the formation of melt accumulations. Pvor<1.5*Prück or also Pvor<2*Prück or also 0.2*Prück<Pvor<0.8*Prück often still applies, preferably Pvor=0.5*Prück. During the backward movement, the maximum laser power is typically reached during the cyclic longitudinal movement of at least the partial laser spot. Note that Pvor>Prück or Pvor=Prück can also be selected as an alternative.

A further development is also advantageous in which a frequency fQ of the transverse movement component and a frequency fL of the longitudinal movement component are the same. It is therefore fQ/fL = 1. The same frequencies fQ and fL can be implemented particularly easily in practice. In addition, the laser welding can be performed particularly smoothly.

A sub-variant of this further development is also preferred, in which the at least one partial laser spot of the at least one partial beam rotates cyclically in time around a circle center by superimposing the transverse movement component and the longitudinal movement component of the at least one partial beam in the plane of the surface of the plate parts is moved, whereby the center of the circle is moved along the welding curve with the feed rate VS. The time cyclical circular movement around the center of the circle is easy to apply and enables uniform welding while optimizing the melt flow and the dynamics of the weld pool. An alternative further development in which a frequency fQ of the transverse movement component and a frequency fL of the longitudinal movement component are unequal is also advantageous. As a result, a large number of setting options can be used for different welding situations and the melt flow and weld pool dynamics can be optimized according to the welding situation. It is preferably the case that fQ/fL=N/M, with N: a natural number and M: a natural number, in particular one of the numbers M or N being one.

Also preferred is a sub-variant of this further development in which the at least one partial laser spot of the at least one partial beam is chronologically cyclic by superimposing the transverse movement component and the longitudinal movement component of the at least one partial beam in the plane of the surface of the plate parts

- is moved in a zigzag fashion around a zigzag center, the zigzag center being moved along the welding curve at the feed rate VS, or

- is moved along a horizontal figure eight, which is moved along the welding curve with the feed speed VS.

These movement patterns have proven themselves in practice for suppressing the formation of melt accumulations. Typically, with the figure of eight, the connection point of the two circles that form the figure of eight and the points of the two circles opposite this connection point are on the welding curve.

Breathing laser spots

Equally advantageous is a variant which provides that, in order to vary the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool, at least one partial laser spot of at least one partial beam in the plane of the surface of the plate parts is cyclically enlarged and reduced over time. This allows the movement in the melt pool to be easily influenced in order to avoid humping. If necessary, also the seam width or an instantaneous size of a vapor capillary can be influenced.

A further development of this variant is preferred, in which, in order to enlarge and reduce at least the partial laser spot, a focus position of at least the partial beam is changed cyclically in time perpendicular to the plane of the surface of the plate parts. This is particularly easy to carry out in practice and leads to good results.

Also preferred is a further development in which it is provided that, in order to vary the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool, a laser spot of the entire laser beam in the plane of the surface of the plate parts is cyclically enlarged and reduced over time. The temporal enlargement and reduction of the laser spot of the entire laser beam is technically particularly easy to implement. To enlarge and reduce the laser spot, a focus position of the entire laser beam can then be changed cyclically in time perpendicularly to the plane of the surface of the plate parts.

Choice of variation frequency

A variant is also advantageous, which is characterized in that the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool is varied cyclically over time with a frequency fM, fM being chosen such that for a normalized frequency NF, the following applies: NF =fM*dw/VS with 0.l<NF<5, with dw: largest diameter of the laser beam on the surface. If fM is selected in such a way that the specified range for NF is observed, good suppression of melt accumulations can usually be achieved.

A variant is also advantageous in which the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool is varied cyclically with a frequency fM, where fM is chosen such that k*0.9*fH<fM<k*l.l*fH, with fH: humping frequency, with fH= VS/laverage, with laverage: average distance from melt accumulations in the solidified weld seam, if the laser welding would take place without the time-cyclic variation of the laser power, and k: natural number. If the frequency fM is chosen to be similar or the same as the humping frequency fH, the formation of melt accumulations can be suppressed particularly efficiently. k=l is preferred.

A further development of this variant is characterized in that before the laser welding of the bipolar plate, a test plate is laser welded with a test laser beam at the feed rate VS in a preliminary measurement, with two test plate parts for the test plate at least in their material type and mate rial strength correspond to the type of material and the material thickness of the plate parts and the test laser beam corresponds to the laser beam in the time average, but does not apply any time-cyclic variation of the power density distribution in the plane of the surface of the two test plate parts in the area of the melt pool, and that the average distance laverage from melt accumulations is determined in a solidified test weld seam of the test plate. In this way, the humping frequency can be easily and accurately determined experimentally. Typically, the test plate (or the test plate parts) is structurally identical to the bipolar plate (or the plate parts).

Also advantageous is a variant in which the power density distribution in the plane of the surface of the two plate parts in the area of the melt pool is varied cyclically over time with a frequency fM, fM being chosen such that 0.8*fnat<fM<1.2* fnat, with fnat: natural weld pool frequency, with fnat= VS/(2*L), with L: weld pool length in welding direction SR. A correspondingly selected (modulation) frequency fM also results in a very efficient suppression of melt accumulations.

Various variants, in particular beam shaping In an advantageous variant, it is further provided that the laser beam is in the form of a superimposed laser beam, comprising at least two partial beams that represent superimposed partial laser beams, the superimposed partial laser beams lying one inside the other on the surface. Typically, a lower local laser power density is established in a radially outer part of the heterodyne laser beam and a locally higher laser power density is established in a radially inner part of the heterodyne laser beam. This allows the vapor capillary to be stabilized during deep penetration welding and the weld pool dynamics to be reduced. The superimposed laser beam typically has concentric superimposed partial laser beams on the workpiece surface.

A preferred further development of this variant is characterized in that the superimposed laser beam comprises a core beam and a ring beam that surrounds the core beam as superimposed partial laser beams, or that the superimposed laser beam has a larger superimposed partial laser beam and a smaller superimposed partial laser beam on the workpiece surface lies within the larger superimposed partial laser beam. This procedure is particularly simple and has proven itself in practice. A heterodyne laser beam with core beam and ring beam is typically generated with a 2-in-1 fiber, and the core beam and ring beam have a common optical axis.

Also preferred is a variant which provides that the plate parts each have a sheet metal thickness BLD of between 50 μm and 150 μm. The plate parts are preferably made of stainless steel, for example type 1.4404. The sheet thickness BLD is preferably 75 μm. Corresponding metallic plate parts can be produced inexpensively and are particularly well suited to the requirements in a bipolar plate of a fuel cell, in particular with regard to corrosion resistance, electrical conductivity and workability during laser welding. Sheet thicknesses between 50 μm and 150 μm combine sufficient robustness with a light and material-saving construction. Also preferred is a variant in which the at least one weld seam comprises one or more self-contained weld seams. Closed weld seams are generally used to seal off fluids in the fuel cell (cooling water, reaction water, reaction gases such as hydrogen, oxygen). Within the scope of the invention, closed weld seams can be manufactured with improved tightness, which makes the invention particularly advantageous here.

A further advantageous variant provides that the at least one weld seam comprises at least one closed weld seam running around the outside of the plate parts. The closed weld seam running around the outside ensures in particular that no coolant (cooling water) escapes into the reaction chambers of a fuel cell stack, and no reaction gases (such as hydrogen and oxygen) can get between the plate parts or even mix in an uncontrolled manner. Accordingly, the improved tightness that can be achieved with the invention is of particular advantage.

A variant is also advantageous which is characterized in that the following applies:

- The laser beam is generated with an infrared laser and has a mean wavelength between 800 nm and 1200 nm, preferably 1030 nm or 1070 nm, or the laser beam is generated with a VIS laser with a mean wavelength between 400 nm and 450 nm or between 500 nm and 550 nm generated; and or

- The beam parameter product SPP of the laser beam is between 0.38 mm*mrad and 16 mm*mrad, preferably with SPP<0.6 mm*mrad in single mode or with SP<3mm*rad in multi mode; and or

- The beam diameter dwl of the partial beam used first in the welding direction on the workpiece is between 10 pm and 300 pm, preferably with 30 pm<dwl<70 pm in single mode or 50 pm<dwl<170 pm in the Multi mode, and the beam diameter dwx of all possible other partial beams is selected with 0.1*dwl<dwx<10*dwl, preferably with dwl=dwx; and or

- The laser power P of the laser beam is between 10W and 2000W, preferably with 50W<P<700W; and or

- The feed rate VS of the laser beam is between 100 mm/s and 5000 mm/s, preferably with 300 mm/s<VS<2000 mm/s; and or

- An imaging ratio AV of a laser lens with which the laser beam is imaged onto the workpiece is between 1:1 and 5:1, preferably with 1.5:1<AV<2:1.

These parameters have proven themselves in practice for the production of the bipolar plates according to the invention.

Furthermore, within the scope of the present invention is a bipolar plate for a fuel cell, produced by welding two plate parts together according to a method according to the invention as described above. Bipolar plates manufactured in this way are characterized by good fluid tightness at the at least one weld seam and achieve a good mechanical and electrical connection between the plate parts. In addition, the bipolar plates can be manufactured particularly quickly.

Further advantages of the invention result from the description and the drawing. Likewise, the features mentioned above and those detailed below can be used according to the invention individually or collectively in any combination. The embodiments shown and described are not to be understood as an exhaustive list, but rather have an exemplary character for the description of the invention. Detailed

the and

The invention is illustrated in the drawing and is explained in more detail using exemplary embodiments.

1 shows a schematic top view of a bipolar plate according to the invention with two plate parts which are connected to one another by a plurality of welded seams which are closed all the way around and a plurality of welded seams which extend in a straight line;

2a shows a schematic top view of a molten pool that is produced with a laser beam according to a first variant of the invention and an associated diagram in which the power density distribution on the surface of the molten pool along axis A at different points in time is plotted;

FIG. 2b shows a diagram of the time-varying laser power of the laser beam from FIG. 2a for carrying out the method according to the invention in accordance with the first variant;

3a shows a diagram of the time-varying laser power of a laser beam for carrying out the method according to the invention according to a second variant;

3b shows a diagram of the time-varying laser power of a laser beam for carrying out the method according to the invention according to a third variant;

4 shows a diagram of the time-varying laser power of a laser beam for carrying out the method according to the invention according to a fourth variant; 5 shows a diagram of the time-varying laser power of two partial beams of a laser beam with the total power of the laser beam remaining the same for carrying out the method according to the invention according to a fifth variant;

6a shows a schematic plan view of a molten bath that is produced with a laser beam according to a sixth variant of the invention, with a time-cyclic shift of the laser beam along the welding direction SR. according to a longitudinal movement component and an associated diagram in which the power density distribution on the surface of the molten pool along the axis A is plotted at different points in time;

FIG. 6b shows a diagram of the time-varying position of the laser beam from FIG. 2a for carrying out the method according to the invention in accordance with the sixth variant;

7a shows a schematic top view of a molten pool that is produced with a laser beam according to a seventh variant of the invention, with a time-cyclic displacement of the laser beam transverse to the welding direction SR according to a transverse movement component;

7b shows a schematic plan view of a molten bath that is produced with a laser beam according to an eighth variant of the invention, with a time-cyclic displacement of the laser beam parallel and transverse to the welding direction SR according to a longitudinal movement component and a transverse movement component;

8a shows a schematic plan view of a molten bath that is produced with a laser beam according to a ninth variant of the invention, with the laser beam being shifted along it cyclically over time a circular path according to a transverse component of movement and a longitudinal component of movement having an equal frequency;

Fig. 8b shows a schematic plan view of a molten bath that is produced with a laser beam according to a tenth variant of the invention, with a time-cyclical displacement of the laser beam along an 8-shaped path according to a transverse movement component and a longitudinal movement component, which is an unequal have frequency;

9 shows a schematic plan view of a molten bath that is produced with a laser beam according to an eleventh variant of the invention, with different sizes of the laser spot produced by the laser beam at different points in time;

10 shows a schematic longitudinal section of two plate parts which are welded by means of a laser beam as shown in FIG. 9 according to the eleventh variant, the laser beam having different focus positions in the direction perpendicular to the workpiece surface at different points in time;

11 shows a schematic plan view of a molten bath that is produced with a laser beam according to the invention, and the laser spot produced by the laser beam to explain parameters for determining the frequency fM;

12 shows a schematic longitudinal section of two test plate parts which, within the scope of the invention, are processed in a preliminary measurement with a test laser beam and in which accumulations of melt form on the surface of the solidified weld seam, for determining the frequency fM; 13 shows a schematic longitudinal section of two plate parts which are welded to one another with a laser beam according to the method according to the invention;

14 shows a schematic view of a melt pool that is produced with a laser beam according to a twelfth variant of the invention, the laser beam being designed as a superimposed laser beam with two partial beams;

15a shows a schematic cross section through a welded seam of a bipolar plate produced according to the method according to the invention, produced with a laser beam with a transverse movement component;

15b shows a schematic cross section through a welded seam of a bipolar plate produced according to the method according to the invention, produced with a laser beam with a transverse movement component and synchronous power modulation.

1 shows, in a schematic top view, a bipolar plate 1 according to the invention for a fuel cell not shown in detail here; the bipolar plate 1 shown was produced within the scope of a method according to the invention.

The bipolar plate 1 is made up of an upper plate part 1a and a lower plate part 1b. The two plate parts la, lb of the bipolar plate 1 are arranged one above the other. The two plate parts la, lb have a profile (not shown). The profiling forms a system (e.g. a meandering or double meandering system) of different channels. The channels between the two plate parts la, lb are cooling fluid channels (typically for cooling water). The channels on the outer surfaces of the two plate parts la, lb are ducts for gas (such as oxygen or hydrogen) and water (which occurs as reaction water in the fuel cell). The two Plate parts la, lb are made of a metallic material such. B. stainless steel. A sheet metal thickness BLD of the plate parts 1a, 1b is 75 μm here in each case; Sheet thicknesses between 50 μm and 150 μm are generally preferred.

The two plate parts 1 a , 1 b are connected to one another by a large number of weld seams 2 (after application of the method according to the invention). The weld seams 2 are shown schematically in dashed lines based on their center lines. A circumferentially closed weld seam 2a runs on the outside of the two plate parts 1a, 1b. Two closed weld seams 2b run around two openings 3, which extend through the bipolar plate 1. Several open (here rectilinear) weld seams 2c also run on the bipolar plate 1. The weld seams 2a, 2b are designed to be fluid-tight, in particular gas-tight. The weld seams 2c are used for the mechanical and electrically highly conductive connection between the two plate parts 1a, 1b. The present inventive method was applied to all welds 2a, 2b, 2c.

2a and 2b show a first variant of the method according to the invention, which is applied to the plate parts of a bipolar plate (as explained in FIG. 1). The upper partial figure of FIG. 2a shows a section of a surface 4 on the upper side of the plate part 1a, on which a laser beam 5 illuminates a laser spot 6 and a molten pool 7 is produced. The lower partial figure of FIG. 2a shows a diagram of a power density distribution LDV of laser power in the area of the melt pool 7 along an axis A on the surface 4, which is caused by the laser beam 5 at different points in time. 2b shows a diagram of the change over time in the laser power P of the laser beam 5.

In the upper partial figure of FIG. 2a, the weld seam 2 being produced is shown schematically in dashed lines on the surface 4 using its center lines.

In the variant shown in FIG. 2a, the weld seam 2 is produced by means of the laser beam 5, which is directed from above (perpendicular to the plane of the drawing) onto the upper plate part 1a. The laser beam 5 generates on the surface 4 of the plate part la the laser spot 6. The laser beam 5 generates the melt pool 7 of melted plate material around the laser spot 6. The melt pool 7 is widest in the area of the laser spot 6 . Contrary to the welding direction SR. the molten pool 7 gradually becomes narrower. The melt pool 7 extends between a front end VE and a rear end HE.

The laser spot 6 is moved along a welding curve 8 as part of a welding feed. In the figure shown here, the welding curve 8 is identical to the center line of the weld seam 2. The laser beam 5 (and thus the laser spot 6) is moved along the welding curve 8 with a basic movement component GBK and a feed rate VS. The direction of feed is marked with an arrow; the feed is along a welding direction SR. In the variant shown, no further movement components are used.

2a and 2b illustrate a first possibility of how the power density distribution LDV in the plane of the laser-facing surface 4 of the plate part la in the area of the melt pool 7 can be varied cyclically over time, with the laser power of the laser beam 5 at a constant size of the laser spot 6 is varied. The diagram of the power density distribution LDV in the region of the molten pool 7 along the axis A on the surface 4 is shown in the lower partial figure of FIG. 2a. The diagram shows the local power density LD in % as a function of the position along the axis A in the plane of the surface 4. The power density LD is greatest at the position where the center 9 of the laser spot 6 lies and falls from the center 9 of the laser spot 6 to the edge 10 of the laser spot 6 down.

Power density distributions LDV are entered in the diagram at three different points in time t1, t2, t3. At time t1, the power density distribution LDV (solid line) in the area of the melt pool 7 has its greatest values and at the position in the center 9 of the laser spot 6 the greatest power density LD, which is normalized to 100% here, is reached. At time t2, the power density distribution LDV (dotted line) in the area of Melt pool 7 smaller values compared to the time tl and at the position in the center 9 of the laser spot 6, a power density LD of about 85% is achieved here. At time t3 the power density distribution LDV (dashed line) in the area of the melt pool 7 has its lowest values compared to the times t1 and t2 and at the position in the center 9 of the laser spot 6 a power density LD of around 70% is reached here.

These power densities LD at the different points in time t1, t2, t3 correlate with the (total) laser power P of the laser beam 5; the laser power P is plotted in the diagram of FIG. 2b as a function of time t. The laser power P of the laser beam 5 is varied sinusoidally in the variant shown here. At time t1, the laser power P of the laser beam 5 reaches its maximum of 100%. The power density distribution LDV in the area of the melt pool 7 also has its greatest values at the time t1. At time t2 the laser power P of the laser beam 5 is approximately 85% and at the time t3 the laser power P of the laser beam 5 is approximately 70%. The power density distribution LDV in the area of the melt pool 7 at the times t2 and t3 also changes in accordance with the laser power P at the times t2 and t3.

By varying the laser power P of the laser beam 5 cyclically over time, the power density distribution LDV in the area of the melt pool 7 can be varied in a simple manner and the formation of humps can be suppressed.

In the variant shown here, the laser beam 5 includes only a partial beam. According to the invention, the laser beam 5 can comprise a partial beam (as shown here) or several partial beams (not shown in detail, but see FIG. 5 or FIG. 14).

3a shows a second variant of the method according to the invention, in which the power density distribution in the plane of the surface of the two plate parts facing the laser in the area of the melt pool is varied by a time-cyclic variation of the laser power P of the laser beam 5. In the diagram shown the (total) laser power P of the laser beam 5 is plotted as a function of time t.

The laser power P of the laser beam 5 is varied in the form of a rectangle in the variant shown here. There are periods when the laser power P is 100% and periods when the laser power P is about 30% here. At the transition between the laser powers P, there is a jump in the laser power P of the laser beam 5 in each case.

3b shows a third variant of the method according to the invention, in which the power density distribution in the plane of the surface of the two plate parts facing the laser in the area of the melt pool is varied by a time-cyclic variation of the laser power P of the laser beam 5. In the diagram shown, the laser power P is plotted as a function of time t.

The laser power P is varied triangularly in the variant shown here. The maximum laser power P is 100% and the lowest laser power P is about 40% here. In between, the laser power P is changed linearly with time t.

4 shows a fourth variant of the method according to the invention, in which the power density distribution in the plane of the surface of the two plate parts facing the laser in the area of the melt pool is varied by a time-cyclic variation of the laser power P of the laser beam 5 . In the diagram shown, the laser power P is plotted as a function of time t.

The variation of the laser power P of the laser beam 5 in the variant shown here takes place as a combination of sinusoidal and triangular changes in the laser power P as a function of the time t. In the area of power peaks (laser power P between approx. 80% and 100%), the laser power P varies triangularly (with a linear increase and a linear drop in laser power), in the area of power troughs (laser power P between approx. 40% and approx. 60%), the laser power P varies sinusoidally and power jumps are set up at transition times tue between the triangular and sinusoidal variation (jump in the laser power P from approx. 60% to approx. 80% and vice versa).

5 shows a fifth variant of the method according to the invention, in which the power density distribution in the plane of the laser-facing surface of the two plate parts in the area of the melt pool is varied by a time-cyclic variation of a power distribution between two partial beams of a laser beam. The partial beams can be set up, for example, as a core beam and ring beam of a superimposed laser beam (cf. FIG. 14 with partial beams 11a, 11b). In the diagram shown, the laser power is plotted as a function of time t.

In the variant shown here, a total power Pges of the laser beam is kept constant at 100%. The first partial beam 11a (cf. FIG. 14) of the laser beam has a partial beam power Plla and the second partial beam 11b (cf. FIG. 14b) of the laser beam has a partial beam power Pllb. The partial beam powers Plla, Pllb of the partial beams of the laser beam result in the total power Pges of the laser beam. The power distribution between the two partial beams is varied by varying the partial beam powers P11a, P1lb in a triangular manner over time. The changes in the partial beam powers P11a, P1lb are of the same magnitude at all times with opposite signs.

Referred to the total power Pges, at time tl the partial beam power Plla is 75% and the partial beam power Pllb is 25%, at time t2 the partial beam powers Plla, Pllb are each 50% and at time t3 the partial beam power Plla is 25% and the partial beam power Pllb is 75% .

In the variant shown (see also FIG. 14 in this regard), the partial beams are partial beams lying one inside the other with a core beam and a core beam Ring beam that were generated, for example, by means of a multiclad fiber (not shown in detail). Alternatively, for example, partial beams lying next to one another can also be involved (not shown in more detail).

6a and 6b shows a sixth variant of the method according to the invention, which is applied to the plate parts of a bipolar plate. The upper partial figure of FIG. 6a shows a section of the surface 4 on the upper side of the plate part 1a, on which the laser beam 5 illuminates the laser spot 6 and the molten pool 7 is produced. The lower partial figure of FIG. 6a shows a diagram of the power density distribution LDV of laser power in the area of the melt pool 7 along the axis A on the surface 4, which is caused by the laser beam 5. 6b shows a diagram of the change over time in a position of the center 9 of the laser spot 6 on the axis A in the reference system of the molten pool 7 advancing with the basic movement component. The molten pool

7 extends between the front end VE and the rear end HE.

In the upper sub-figure of FIG. 6a, the weld seam 2 being produced is shown schematically in dashed lines using its center lines on the surface 4.

In the variant shown in FIG. 6, the weld seam 2 is produced by means of the laser beam 5, which is directed from above (perpendicular to the plane of the drawing) onto the upper plate part 1a. The laser beam 5 generates the laser spot 6 on the surface 4 of the plate part 1a. The laser beam 5 generates the molten bath 7 of melted plate material around the laser spot 6. The melt pool 7 is widest in the area of the laser spot 6 . Contrary to the welding direction SR. the molten pool 7 gradually becomes narrower.

The laser spot 6 is within the scope of a welding feed on the welding curve

8 moves. In the figure shown here, the welding curve 8 is identical to the center line of the weld seam 2. The laser beam 5, and thus also the laser spot 6, are moved along the welding curve 8 with the basic movement components GBK and the feed rate VS. The direction of feed is marked with an arrow; the feed runs along the welding direction SR.

Fig. 6a and Fig. 6b illustrate a further possibility of subjecting the power density distribution LDV in the plane of the surface 4 of the plate part 1a facing the laser in the area of the melt pool 7 to a temporally cyclical variation, the basic movement component GBK of the laser beam 5 being a longitudinal Be - Movement component LBK is superimposed with a frequency fL; the laser beam 5 is reciprocated along the welding direction SR. In the upper partial figure of FIG. 6a, the laser spot 6 is shown in the region of the melt pool 7 at three different positions PI, P2, P3. The marked positions PI, P2, P3 each correspond to the center 9 of the laser spot 6 at different points in time t1, t2, t3. The positions PI, P2, P3 are specified in the reference system of the melt pool 7 moving with the basic movement component GBK. The melt pool 7 extends between the front end VE and the rear end HE. The positions PI, P2, P3 are all on the welding curve 8. To illustrate the longitudinal movement, the laser spot 6 is shown as a broken line at position PI, as a continuous line at position P2 and as a dot at position 3.

The diagram of the power density distribution LDV in the region of the molten pool 7 along the axis A on the surface 4 is shown in the lower partial figure of FIG. 6a. The diagram shows the local power density LD in % as a function of the position along the axis A in the plane of the surface 4. The power density LD is greatest at the position where the center 9 of the laser spot 6 lies and falls from the center 9 of the laser spot 6 to the edge 10 of the laser spot 6 down.

Power density distributions LDV are entered in the diagram at four different points in time t1, t2, t3, t4. At time t1, the power density distribution LDV in the area of the melt pool 7 (solid line) has its greatest values at position P2. At time t2, the power density distribution LDV in the area of the melt pool 7 (dashed line) has its greatest values at position PI. At time t3, the power density distribution in the area of the melt pool 7 (solid line) their largest values again at position P2. At time t4, the power density distribution in the area of the melt pool 7 (dotted line) has its greatest values at position P3.

The diagram of FIG. 6b shows the position of the center 9 of the laser spot 6 on the axis A as a function of the time t. The position of the center 9 is varied sinusoidally in the variant shown here. At the point in time t1, the center 9 of the laser spot 6 is in position P2. At time t2, the center 9 of the laser spot 6 is in position PI. At time t3, the center 9 of the laser spot 6 is back in position P2. At time t4, the center 9 of laser spot 6 is in position 3. The power density distribution LDV in the area of the melt pool 7 at times t1, t2, t3, t4 also changes according to the position at times t1, t2, t3, t4.

The power density distribution LDV in the region of the melt pool 7 can be varied in a simple manner and the formation of humps can be suppressed by the time-cyclic variation of the location of the laser spot according to the longitudinal movement components LBK.

7a explains a seventh variant of the method according to the invention, in which the power density distribution in the plane of the surface of the two plate parts facing the laser in the area of the melt pool 7 is varied by a time-cyclic variation of the position of the laser beam 5.

The laser beam 5 is with the basic movement components GBK with the feed rate VS along the welding curve 8 in the welding direction SR. emotional. In addition, a transverse movement component QBK with a frequency fQ is superimposed on the laser beam 5; the transverse movement of the laser beam 5 takes place perpendicularly to the welding curve 8 back and forth.

In FIG. 7a, the laser spot 6 is shown in the area of the melt pool 7 at the three positions P1, P2, P3. The marked positions PI, P2, P3 correspond to the center 9 of the laser spot 6 at different points in time. Position P2 is on Welding curve 8, the positions PI and P3 are each shifted transversely to the welding curve 8. To illustrate the transverse movement and the different positions PI, P2, P3 to which the center 9 of the laser spot 6 is shifted, the laser spot is shown as a dotted circle at position PI, as a solid circle at position P2 and as a dashed circle at position P3 .

7b explains an eighth variant of the method according to the invention, in which the power density distribution in the plane of the surface of the two plate parts facing the laser in the area of the melt pool 7 is varied by a time-cyclic variation of the position of the laser beam 5.

The laser beam 5 is with the basic movement components GBK with the feed rate VS along the welding curve 8 in the welding direction SR. emotional. In addition, the transverse movement component QBK with the frequency fQ and the longitudinal movement component LBK with the frequency fL are superimposed on the laser beam 5; the transverse movement of the laser beam 5 takes place perpendicular to the welding curve 8 back and forth and the longitudinal movement of the laser beam 5 takes place parallel to the welding curve 8 back and forth.

In FIG. 7b, the laser spot 6 is shown in the area of the melt pool 7 at the three positions PI, P2, P3. The marked positions PI, P2, P3 correspond to the center 9 of the laser spot 6 at different points in time. Position P2 is on welding curve 8, positions PI and P3 are shifted transversely and longitudinally to welding curve 8, respectively. To illustrate the transverse movement and the different positions PI, P2, P3 to which the center 9 of the laser spot 6 is shifted, the laser spot is shown as a dotted circle at position PI, as a solid circle at position P2 and as a dashed circle at position P3 shown.

FIG. 8a shows a ninth variant of the method according to the invention, in which, similar to the eighth variant from FIG. 7b, both a longitudinal movement component and a transverse movement component are used. The basic movement component GBK with the feed rate VS is applied to the laser beam 5 . In addition, the transverse movement component with the frequency fQ and the longitudinal movement component with the frequency fL are superimposed on the laser beam 5, which are the same here (fL=fQ). The longitudinal movement components and the transverse movement components are each sinusoidal with the same amplitude and are phase-shifted by 90° with respect to one another. This results in a circle (circular path) 12, which the laser beam 5 or the associated laser spot 6 travels cyclically over time with its center (spot center) on the surface. The circle 12 has a circle center 13. The circle center 13 is guided along the welding curve 8 at the feed rate VS.

FIG. 8b shows a tenth variant of the method according to the invention, in which, similar to the eighth variant from FIG. 7b, both a longitudinal movement component and a transverse movement component are used.

The basic movement component GBK with the feed rate VS is applied to the laser beam 5 . In addition, the transverse movement component with the frequency fQ and the longitudinal movement component with the frequency fL are superimposed on the laser beam 5, which are unequal here (fL^fQ). The frequency fQ of the transverse movement component is twice the frequency fL of the longitudinal movement component. Furthermore, the amplitude of the longitudinal sinusoidal motion component is twice the amplitude of the transverse sinusoidal motion components; the phase shift is zero. This results in two circles 12a, 12b lying next to one another (in the direction of the basic movement components GBK), which form a figure eight (8-shaped path) 14 and which the laser spot 6 with its spot center on the surface cyclically runs through in terms of time. The movement of the laser spot 6 along the lying eight (8-shaped path) 14 is indicated by a directional arrow RP. The horizontal eight 14 moves overall with the feed speed VS along the welding curve 8. An eleventh variant of the method according to the invention is explained in FIG.

The laser beam 5 is moved along the welding curve 8 with the basic movement component GBK at the feed rate VS. In addition, the laser spot 6 of the laser beam 5 is cyclically enlarged and reduced over time. The laser spot 6 is shown here once with a solid line (at time t1) and once with a dotted line (at time t2). This cyclical enlargement and reduction of the laser spot 6 over time also varies the power density distribution in the plane of the laser-facing surface of the two plate parts in the area of the melt pool 7 cyclically, which means that the formation of humps can be suppressed. In the eleventh variant presented here, the (entire) laser power of the laser beam 5 remains constant. If one considers the power density distribution in the area of the molten pool 7 along the axis A on the surface in this example (according to Fig. 2a or Fig. 6a), then at time tl the power density distribution in the area of the molten pool 7 is narrower with higher values for the power density and at time t2 the power density distribution in the area of the molten pool 7 is broader with lower values for the power densities (not shown in more detail).

In FIG. 10, the eleventh variant of the method according to the invention from FIG.

The plate parts la, lb are arranged one above the other. The laser beam 5 generates the laser spot 6 in a plane E of the surface 4 of the first plate part la. emotional. The laser beam 5 generates the molten pool 7 of melted plate material around the laser spot 6 . In the wake of the molten bath 7, the melted plate material solidifies and forms the finished weld seam 2. In order to suppress the formation of humps (as already explained in FIG. 9 in a top view of the melt pool 7), the laser spot 6 in the plane E is cyclically reduced and enlarged over time. In the variant shown here, the reduction and enlargement is brought about by a focus position (focus plane/location of the smallest beam diameter) FP1, FP2 of the laser beam 5 being chronologically shifted perpendicularly to the plane E. The laser beam 5 is shown in FIG. 10 at the time t1 (solid lines of the laser beam 5) and at the time t2 (dotted lines of the laser beam 5).

At the time tl, the focus position FP1 is in the plane E. The laser spot 6 on the melt pool 7 has a diameter dl. At time t2, the focus position FP2 of the laser beam 5 was shifted by a height H perpendicular to the plane E in the direction of the plate parts 1a, 1b. The laser spot 6 on the molten bath then has a diameter d2 that is larger than the diameter dl of the laser spot 6 at the time t1.

It should be noted that a developing vapor capillary is not shown in detail in FIG. 10 for the sake of simplicity.

Possibilities are shown in FIG. 11 with which the frequency fM of a temporally cyclical variation of the power density distribution in the plane of the surface of the two plate parts facing the laser in the area of the melt pool 7 can be selected. The melt pool 7, which is generated with the laser beam 5 and its laser spot 6, is shown in a schematic top view.

The (here circular) laser spot 6 has a (largest) diameter dw of 50 pm here. The size of the laser spot 6 can be determined using the 86% criterion. For this purpose, a circular area is placed around the center of the intensity profile of the laser beam with a diameter such that 86% of the laser power lies within the circular area. The laser beam 5 is with the basic movement components GBK are moved along the welding curve 8 with the feed rate VS of 1000mm/s here. The molten pool 7 has a length L along the welding direction SR, L being approximately 60 μm in the selected example

On the one hand, for a normalized frequency NF:

NF = fM*dw/VS.

Typically, the normalized frequency NF is between a value of 0.1 and 5. Converting the above equation to fM gives the following expression: fM = NF*VS/dw.

If the desired normalized frequency NF (or a desired interval for NF) is specified and the largest diameter dw and the feed rate VS are known, the required frequency fM (or a required interval for fM) can be estimated.

For example, assuming that NF = 1, the rearranged equation for fM in the present example looks like this: fM = l*(1000mm/s)/(50pm) = 20kHz.

If an interval of 0.1 to 5 is allowed for NF, then in the example chosen fM can be selected in an interval of 2kHz to 100kHz.

On the other hand, a natural melt pool frequency fnat can be determined. Typically, the frequency fM should be in a range of 0.8*fnat and 1.2*fnat. The natural melt pool frequency fnat can be determined using the following equation: fnat = VS/(2*L). If the feed rate VS and the weld pool length L are known, the natural weld pool frequency fnat and, as a result, the frequency fM can be determined. In the selected example, the above equation for fnat results in: fnat = (1000mm/s)/(2*60pm) = 8.33kHz.

The power density distribution in the plane of the laser-facing surface of the two plate parts in the area of the melt pool 7 should therefore be varied cyclically over time with a frequency fM with a value between 6.67 kHz and 10 kHz.

FIG. 12 shows how the frequency fM can be selected with the aid of a test plate 15 and a preliminary measurement. The test panel 15 is shown in a schematic longitudinal section.

To produce the test panel 15, two test panel parts 15a, 15b are arranged one above the other. In the variant shown, the test plate 15 is structurally identical to the bipolar plate, or the test plate parts 15a, 15b are structurally identical to the plate parts for the bipolar plate; In particular, the test plate ("test bipolar plate") 15 is made of the same material as the finished bipolar plate is to be made of, and the material thicknesses (plate thicknesses) 26a, 26b of the test plate parts 15a, 15b correspond to the material thicknesses (plate thicknesses ) of the plate parts of the bipolar plate to be manufactured (cf. Bzz. 27a, 27b in Fig. 13). is to be used for the production of the bipolar plates, but without using a temporally cyclic variation of the power density distribution in the area of the melt pool 7. The test laser beam 16 generates a test laser spot 17 in the plane E of the surface 4 of the first test plate part 15a Test laser beam 16 is moved with the basic movement component GBK at the feed rate VS, here 1000 mm/s. The test laser beam 16 generates the molten pool 7 of melted plate material around the test laser spot 17 . The finished test weld seam 25 forms in the wake of the molten bath 7. Melt accumulations ("humps") 18 form on the test weld seam 25. These melt accumulations 18 occur at regular intervals II, 12, 13 at constant feed rate VS. Off From these distances II, 12, 13 an average distance laverage can then be determined from melt accumulations 18. Here the melt accumulations 18 are all spaced apart by approximately 50 pm, the average distance laverage is then approximately 50 pm.

The frequency fM can be determined based on the preliminary measurement of the test plate 15 as follows:

A humping frequency fH can be determined with the preliminary measurement. Typically, the frequency fM should be in a range of 0.9*k*fH and 1.1*k*fH, where k is a natural number. The humping frequency fH can be determined using the following equation: fH = VS/laverage.

If the feed rate VS and the mean distance laverage are known, the humping frequency fH and, as a result, the frequency fM can be determined. In the selected example, the above equation for fH : fH = (1000mm/s)/(50pm) = 20kHz.

The power density distribution in the plane of the laser-facing surface of the two plate parts in the area of the melt pool 7 of the bipolar plate to be manufactured should therefore be varied cyclically over time with a frequency fM with a value between k*18 kHz and k*22 kHz. With k=l, for example, fM should then be selected between 18kHz and 22kHz. FIG. 13 shows a schematic longitudinal section of a bipolar plate 1 which is processed according to the first variant (FIGS. 2a, 2b) of the method according to the invention.

Before the actual welding, a preliminary measurement was made as described in Fig. 12 and a humping frequency fH was determined, which in turn can be used to determine the frequency fM with which the power density distribution in the plane of the laser-facing surface 4 of the plate part la in the area of the melt pool 7 over time must be varied cyclically in order to suppress the formation of melt accumulations.

The plate parts la, lb are arranged one above the other. The laser beam 5 is moved with the basic movement component GBK at the feed rate VS. The laser beam 5 generates the melt pool 7 of melted plate material. Its power density distribution on the surface 4 in the area of the melt pool 7 is varied cyclically over time with the frequency fM, whereby melt accumulations are suppressed. In the wake of the molten bath 7, the finished weld seam 2 is formed without any accumulations of melt.

FIG. 14 shows an alternative laser beam 5 in a schematic plan view of a molten bath 7, as can be used for the method according to the invention. Laser beam 5 is designed here as a superimposed laser beam 19 with the two partial beams 11a, 11b. The partial beams 11a, 11b lie in one another here as superimposed partial laser beams 19a, 19b. The superimposed partial laser beam 19a is designed as a core beam 20 here. The core beam 20 is surrounded in the form of a ring by a superimposed partial laser beam 19b which is in the form of a ring beam 21 . The superimposed laser beam 19 can be generated, for example, by a multiclad fiber (such as a 2in1 fiber).

The core beam 20 and the ring beam 21 produce a laser spot 6 which has a core portion 22 and a ring portion 23 . The ring portion 23 is to Core portion 22 arranged. The superimposed laser beam 19 generates the melt pool 7 and is moved along the welding curve 8 with the basic movement component GBK at the feed rate VS. A temporally cyclical variation of the power density distribution in the plane of the laser-facing surface of the two plate parts in the area of the melt pool 7 can take place, for example, by redistributing the power between the core part 22 and the ring part 23 (distributed back and forth).

Core portion 22 and ring portion 23 are circular here.

FIG. 15a shows a schematic cross section of a bipolar plate 1 which was manufactured according to the seventh variant of the method according to the invention (see FIG. 7a). The cross section is perpendicular to the direction of welding.

The plate parts 1a, 1b are connected to one another by the weld seam 2 and no accumulations of melt have formed on the surface 4. In the variant shown here, the power density distribution on the surface 4 in the area of the molten pool was varied cyclically over time by superimposing a transverse movement component (the left and right back and forth in Fig. 15a) on the basic movement component of the laser beam. As a result, during the welding, the lateral edges of the molten pool with respect to the transverse direction heated up more or the plate material melted more deeply than in the middle. As a result, deformations 24 ("molars") are formed in the solidified melt in the weld seam 2 at the lateral edges; in other words, the weld seam 2 is deeper at the lateral edges in the transverse direction in the region of the deformations 24 than in the middle.

In order to optimize the shape of the weld seam 2, the laser power can also be varied cyclically over time in order to avoid locally stronger heating at the edges of the melt pool and thus the deformations 24 in the weld seam 2. For this purpose, the laser power can be reduced in the times when the laser spot is approaching the lateral edges compared to times when where the laser spot is near the center of the weld pool, or near the weld curve (center line of the weld). A correspondingly welded weld seam is shown in the schematic cross section of a bipolar plate 1 in FIG. 15b. The weld seam 2 has a substantially constant depth in the transverse direction.

1 bipolar plate la upper plate part lb lower plate part

2 weld

2a circumferentially closed weld seam

2b closed weld seam

2c open (straight) weld

3 breakthrough

4 surface (of the upper part of the plate)

5 laser beam

6 laser spot

7 weld pool

8 welding curve

9 center (of the laser spot)

10 edge (of the laser spot)

11a, 11b partial beam

12 circle

13 circle center

14 lying eight

15 test plate

15a upper test plate part

15b lower test plate part

16 test laser beam

17 test laser spot

18 melt accumulations / humps

19 Overlay Laser Beam

19a, 19b superimposed partial laser beam

20 core beam

21 ring ray

22 core portion

23 ring share 24 Deformation 25 Test weld 26a Material thickness of upper test plate part 26b Material thickness of lower test plate part 27a Material thickness of upper plate part 27b Material thickness of lower plate part A axis BLD Sheet thickness dl, d2 diameter (partial) Laser spot dw largest diameter E plane (of the surface of the plate parts) FP1, FP2 focus position GBK basic movement component H height (between the focus positions of the laser beam at different times)

HE rear end of weld pool L weld pool length (in welding direction SR.)

II, 12, 13 Distance between two accumulations of melt LBK Longitudinal movement component LDV Power density distribution P Laser power Pges Total power of the laser beam

Plla, Pllb Partial beam power PI, P2, P3 Position (of the center of the laser spot) QBK Transversal movement component RP Directional arrow SR Welding direction t Time tl-t4 Time tue Transition time VE Front end of the weld pool

Claims

patent claims

1. A method for laser welding a bipolar plate (1) for a fuel cell, in which two plate parts (1a, 1b) are welded to one another along at least one weld seam (2, 2a, 2b, 2c), the laser welding being carried out with a laser beam (5), wherein the laser beam (5) has a basic movement component (GBK) with a feed rate VS in a welding direction (SR.) along a welding curve (8) in a plane (E) of a surface (4) of the two plate parts (la, lb) relative to the plate parts (1a, 1b), and the welding curve (8) runs along the weld seam (2, 2a, 2b, 2c), the laser beam (5) generating a melt pool (7) in the plate parts (1a, 1b). , and the laser beam (5) causes a power density distribution (LDV) of laser radiation in the plane (E) of the surface (4) of the two plate parts (la, lb) in the area of the melt pool (7), and the laser beam (5) produces a or several partial beams (11a, 11b), characterized in that the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (1a, 1b) in the area of the melt pool (7) is varied cyclically over time.

2. The method according to claim 1, characterized in that to vary the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (1a, 1b) in the area of the melt pool (7) a laser power at least one partial beam (11a, 11b) of the laser beam (5) is varied cyclically over time.

3. The method according to claim 2, characterized in that to vary the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (1a, 1b) in the area of the melt pool (7), a power distribution between at least two partial beams (11a, 11b) is varied cyclically over time, with a total power (Pges) of the laser beam (5) being kept constant over time.

4. The method according to any one of the preceding claims, characterized in that to vary the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (1a, 1b) in the area of the melt pool (7) at least one partial beam ( 11a, 11b) in the plane (E) of the surface (4) of the plate parts (la, lb) a transverse movement component (QBK) is superimposed, so that at least one partial laser spot of the at least one partial beam (11a, 11b) is transverse to the Welding curve (8) is moved back and forth cyclically over time.

5. The method according to any one of the preceding claims, characterized in that to vary the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (1a, 1b) in the area of the molten pool (7) at least one partial beam ( 11a, 11b) a longitudinal movement component (LBK) is superimposed in the plane (E) of the surface (4) of the plate parts (la, lb), so that at least one partial laser spot of the at least one partial beam (11a, 11b) is parallel to the Welding curve (8) is moved back and forth cyclically over time.

6. The method according to claims 4 and 5, characterized in that a frequency fQ of the transverse movement component (QBK) and a frequency fL of the longitudinal movement component (LBK) are the same.

7. The method according to claim 6, characterized in that by superimposing the transverse movement component (QBK) and the longitudinal movement component (LBK) of the at least one partial beam (11a, 11b) in the plane (E) of the surface (4) of the plate parts (la, lb) the at least one partial laser spot of the at least one partial beam (11a, 11b) is moved cyclically in a circular manner around a circle center (13), the circle center (13) moving at the feed rate VS along the Welding curve (8) is moved.

8. The method according to claims 4 and 5, characterized in that a frequency fQ of the transverse movement component (QBK) and a frequency fL of the longitudinal movement component (LBK) are unequal.

9. The method according to any one of the preceding claims, characterized in that to vary the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (1a, 1b) in the region of the molten pool (7) at least one part laser spot of at least one partial beam (11a, 11b) in the plane (E) of the surface (4) of the plate parts (la, lb) is cyclically enlarged and reduced over time.

10. The method according to claim 9, characterized in that to enlarge and reduce at least the partial laser spot, a focus position (FP) of at least the partial beam (11a, 11b) perpendicular to the plane (E) of the surface (4) of the plate parts (la, lb ) is changed cyclically over time.

11. The method according to claim 9 or 10, characterized in that in order to vary the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (la, lb) in the region of the melt pool (7), a laser spot (6 ) of the entire laser beam (5) in the plane (E) of the surface (4) of the plate parts (la, lb) is cyclically increased and decreased over time.

12. The method according to any one of the preceding claims, characterized in that the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (la, lb) in the region of the melt pool (7) with a frequency fM cyclically over time is varied, where fM is chosen such that for a normalized frequency NF: NF=fM*dw/VS with 0.1<NF<5, with dw: largest diameter of the laser beam (5) on the surface (4).

13. The method according to any one of the preceding claims, characterized in that the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (la, lb) in the region of the melt pool (7) with a frequency fM cyclically over time is varied, where fM is chosen such that k*0.9*fH<fM<k*l.l*fH, with fH: humping frequency, with fH= VS/laverage, with laverage: average distance from melt accumulations ( 18) in the solidified weld (2, 2a, 2b, 2c) if the laser welding would occur without the time-cyclic variation of the laser power, and k: natural number.

14. The method according to claim 13, characterized in that before the laser welding of the bipolar plate (1) in a preliminary measurement, a test plate (15) is laser welded with a test laser beam (16) at the feed rate VS, with two test plate parts (15a, 15b) for the test plate (15) at least in their material type and material strength (26a, 26b) with the material type and material strength (27a, 27b) of the plate parts (1, lb) match and the test - The laser beam (16) corresponds to the laser beam (5) on average over time, but there is no cyclical variation in the power density distribution (LDV) in the plane (E) of the surface (4) of the two test plate parts (15a, 15b) in the area of the melt pool (7) applies, and that in a solidified test weld seam (25) of the test plate (15) the average distance laverage from melt accumulations (18) is determined.

15. The method according to any one of the preceding claims, characterized in that the power density distribution (LDV) in the plane (E) of the surface (4) of the two plate parts (la, lb) in the region of the melt pool (7) with a frequency fM cyclically over time is varied, where fM is chosen so that 0.8*fnat<fM<l.2*fnat, with fnat: natural melt pool frequency, with fnat= VS/(2*L), with L: melt pool length in welding direction (SR) .

16. The method according to any one of the preceding claims, characterized in that the laser beam (5) is designed as a superimposed laser beam (19) comprising at least two partial beams (11a, 11b) that represent superimposed partial laser beams (19a, 19b), wherein the superimposition partial laser beams (19a, 19b) on the surface (4) are in each other.

17. The method according to any one of the preceding claims, characterized in that the at least one weld seam (2, 2a, 2b, 2c) comprises one or more self-contained weld seams (2a, 2b).

18. bipolar plate (1) for a fuel cell, produced by welding two plate parts (la, lb) according to a method according to any one of the preceding claims.