Laser surface structuring is becoming increasingly important in the industry for tool and mold making. While structured surfaces contribute to minimizing friction in combustion engines or to increasing efficiency of light-emitting diode-based lighting systems, surface texturing is evolving a quality feature of products with regard to optical and haptic properties. Currently used manufacturing processes for tool texturing like photochemical etching are limited in precision and in flexibility. To establish a digital process chain and to increase the design flexibility, laser ablation with (ultra) short pulse laser radiation is becoming an increasingly important technology. In the research project “eVerest,” all necessary parts of a laser texture processing are integrated into the machine and operating concept, e.g., the virtual design of the product including unrolling and visualization of the textures. Finally, new process strategies and advanced machine and system technologies are developed.

Laser surface structuring as a part of materials processing is becoming increasingly important in the industry for tool and mold making. Functionality in the user interface is an increasingly decisive factor for the success and acceptance of a product. This applies to a wide range of products, some of which have very complex requirements for the technical design of surfaces. While structured surfaces contribute to minimize friction in combustion engines or to increase efficiency of light-emitting diode-based lighting systems due to their functionality, surface texturing is evolving a quality feature of products with regard to optical and haptic properties.1–3 In the case of automobiles, this primarily includes the interior, which changes from classic leather grain to technical and functional surface.4 Together with new materials, it is also possible to achieve haptic properties with which an entirely new character of surfaces in the automotive sector can be achieved. Figure 1 shows the process of manufacturing an engine cover. First, a digital mock-up of the component is generated (a). Based on this, a suitable injection molding tool is manufactured and laser-structured (b). Finally, the plastic components are produced by a replication process in large series (c).

FIG. 1.

(a) Digital mock-up/rendering and (b) laser-structured serial tool for the production of the (c) engine cover as the final product.

FIG. 1.

(a) Digital mock-up/rendering and (b) laser-structured serial tool for the production of the (c) engine cover as the final product.

Close modal

Today, these structures are manufactured with complex process chains that, however, allow only very limited flexibility. The current state of the art for the execution of structures is almost exclusively photochemical etching, in which a carrier foil is used to transfer a lacquer structure into the tool, which partially protects against corrosion attack in the subsequent acid bath. In this process, the quality is decisively determined by the manual process of filming and the correct execution of the etching. However, large quantities of etching solutions need to be disposed of during this process.5,6 In addition, the niche of laser texturing has developed in recent years.7–9 Here, the structure is worked out by means of laser ablation. The procedure offers the advantage of a complete digital process chain (precision, quality, and reproducibility) as well as the extension of the grain spectrum by new structures (extension of the design degrees of freedom). Since product design is a crucial sales criterion for future products, the still young technology will be given the greatest opportunity to further increase its value and to develop unique selling propositions in international competition. In this context, only the various functions of a grained surface should be remembered: both aesthetics (appearance, gloss level, and color) and haptics (smooth, rough, technical, and organic) as well as properties in everyday use (cleaning, wear behavior, aging, and sensitivity to scratches) are decisively determined by the structure.

For the production of these tools and structures, the technology of laser ablation is well suited with which microscale structures can be produced. Surface structuring by nanosecond pulsed laser sources is established for industrial use with a high throughput, which, however, allows ablation geometries >50 μm.10,11 Ultrashort laser pulses still have shown their advantages regarding highly precise ablation while minimal thermal influencing the material in diverse applications. However, one of the main drawbacks is the low productivity of the ultrashort pulse laser process.12,13

While in the past the available power was not sufficient to scale productivity, today high power ultrashort pulse laser sources with output powers >100 W are commercially available.14–17 Nevertheless, the challenge is to use this power in an application. If this objective can be solved, a large number of applications will open up not only in automotive industry, but also in printing industry or other consumer goods such as furniture, sports equipment, electronic housing shells, and microstructured surfaces from medical technology.

In the collaborative research project “eVerest,” a consortium of five industrial partners and three research institutes are developing machine and system technology for the efficient production of large-format 3D forming tools that have design surfaces. The project includes players from laser manufacturing through system integration to the automaker all along the entire value chain: Volkswagen, DMG Mori, Scanlab, Precitec, Amphos, Münster University of Applied Sciences, Visual Computing Institute RWTH Aachen University, and Fraunhofer ILT. Table I gives an overview about the project partner's individual tasks.

TABLE I.

Overview about the “eVerest” project partner's individual tasks.

DMG Mori Development and validation of a novel machine technology for precision laser structuring on large components 
Amphos Development of a high power fiber guided laser beam source with a pulse duration of 1 ps 
Scanlab Development of a subsystem for highly dynamic positioning of the laser focus on free-form surfaces 
Precitec Development of an interferometric measurement system for in situ process monitoring (topography, depth measurement) 
Volkswagen Further development of today's laser texture technology in the area of software and hardware and evaluation of new options for automotive interior design. 
Fraunhofer ILT Process investigations in the field of precision ablation with an ultrashort pulse laser source and selective laser polishing 
Münster University of Applied Sciences Development of a fast and novel z-Shifter on the basis of a fast deformable mirror 
Visual Computing Institute, RWTH Aachen University Development of a software solution for user- and component-specific generation of design features on large components 
DMG Mori Development and validation of a novel machine technology for precision laser structuring on large components 
Amphos Development of a high power fiber guided laser beam source with a pulse duration of 1 ps 
Scanlab Development of a subsystem for highly dynamic positioning of the laser focus on free-form surfaces 
Precitec Development of an interferometric measurement system for in situ process monitoring (topography, depth measurement) 
Volkswagen Further development of today's laser texture technology in the area of software and hardware and evaluation of new options for automotive interior design. 
Fraunhofer ILT Process investigations in the field of precision ablation with an ultrashort pulse laser source and selective laser polishing 
Münster University of Applied Sciences Development of a fast and novel z-Shifter on the basis of a fast deformable mirror 
Visual Computing Institute, RWTH Aachen University Development of a software solution for user- and component-specific generation of design features on large components 

The aim of the project is to develop a highly flexible machine technology for precision laser processing, with which functional structures in tools and components can be achieved at highest geometric resolutions in the micrometer range without essential knowledge of the actual technology. Its central goal is to develop highly flexible machine technology for precision laser processing, with which users can achieve the highest geometric resolution in the micrometer range without needing significant knowledge of the actual technology textures in tools (e.g., injection mold) and components (e.g., instrument panel) while also increasing the effective ablation rate. The main topics encompass, among others, the integrative components of innovative systems engineering:

  • Development of a fiber-coupled high power ultrashort pulse laser source

  • Process development for a sequential photonic process chain consisting of laser ablation and laser polishing

  • System technology to increase the ablation rate, including a fast z-shifter and intelligent algorithms for scanner control

  • Virtual component design including handling, synthesis, and visualization of the structures including new possibilities of laser surface processing (high resolution by ultrashort pulse structuring, gloss level adjustment by laser polishing)

  • Optical coherence tomography (OCT) short coherence sensor technology for the analysis of topographic surface profiles

With this highly integrated machine technology, which is novel in terms of data flow and technology integration, the time required to produce a functional surface is to be one third compared to the state of the art. With the “eVerest” project, the consortium is contributing to the BMBF's research program “Innovations for tomorrow's production, services and work.”

Ultrafast laser sources seem to be the ideal candidate for all kind of precise laser material processing tasks. High intensity of several megawatts allows for processing of any material (metal, glass, plastic, semiconductors, etc.), and short pulse duration results in a kind of nonthermal dominated ablation process. As the parts that need to be manufactured cover a broad range of geometry and material, a broad range of laser parameters is necessary to address the different applications. Drilling requires high pulse energy and structuring rather than low pulse energy and high repetition rate, and the burst mode is a special method to improve ablation productivity. In addition, the average power directly translates into processing time. Therefore, the ideal laser has an output power of several 100 W and can be tuned in repetition rate, pulse energy, and pulse duration. InnoSlab amplification concept allows for such flexible lasers due to the quasisingle-pass amplification path (cf. Fig. 2). The architecture is quite simple: a fiber-based seed laser determines the time structure (repetition rate, pulse duration, and burst pattern) and the InnoSlab crystal amplifier adds the power to the multi-100 W range. Due to the large beam diameter on the laser crystal, a pulse energy of several millijoules can be achieved in a compact setup.

FIG. 2.

Illustration of the InnoSlab amplification scheme.

FIG. 2.

Illustration of the InnoSlab amplification scheme.

Close modal

In addition to the laser source a fiber coupling unit for the high power InnoSlab laser has been developed within the project. A transmission of more than 90% at a power level of 135 W for an 8 m fiber has been achieved. The beam profile shows the typical Kagome-type intensity profile with a beam quality of approximately M2 = 1.3 (cf. Fig. 3).

FIG. 3.

Transmission of the peak power through a hollow core fiber with the characteristic Kagome-type intensity beam profile.

FIG. 3.

Transmission of the peak power through a hollow core fiber with the characteristic Kagome-type intensity beam profile.

Close modal

For special tasks, it could be favorable to add a pause in the burst like it is shown in Fig. 4. With these kinds of special burst pattern, the ablation quality can be increased due to the minimized thermal load of the workpiece within the process. This was already shown in Brenner et al. as an intermediate result during the “eVerest” project.18 

FIG. 4.

Examples for the special burst pattern generated from a 40 MHz oscillator: (a) 8 PpB with a pause of 104 ns before the last pulse and (b) 6 PpB with a pause of 104 ns after two pulses.

FIG. 4.

Examples for the special burst pattern generated from a 40 MHz oscillator: (a) 8 PpB with a pause of 104 ns before the last pulse and (b) 6 PpB with a pause of 104 ns after two pulses.

Close modal

The experimental tests for process investigations regarding laser ablation and laser polishing are done by researchers of Fraunhofer ILT. The tests are conducted on hot-working steel blanks (1.2738) that are quite usual in the tool industry. The samples have a grinded initial surface with a roughness of Sa = 0.38 ± 0.02 μm. Squared cavities with a dimension of 5 × 5 mm2 are produced applying a bidirectional scan strategy with an angle of 0° and 90°. For all applied laser ablation processes, the number of repeats (N) is selected so that at the end of the ablation process the cavity has a depth of zabl>30μm in order to keep the measurement error small compared to the overall ablation depth. For the presented results regarding laser polishing, the number of repeats is kept constant at two scans per cavity.

The experimental setup consists of a high power ultrafast laser source from Amphos with a maximum average power of Pav=400W and an adjustable pulse duration τ between 20 and 2 ps. The amplifier is based on the InnoSlab concept, a design of the master oscillator power amplifier that was developed at Fraunhofer ILT (cf. Fig. 2). The emitted laser radiation has a wavelength of 1030 nm (Yb:YAG) and an adjustable pulse repetition rate frep between 400 kHz and 1 MHz for bursts. The seeder frequency fSeed is 39.29 MHz that can be fully utilized for a single pulse. The laser beam is moved by a Scanlab galvanometer scanner (excelliSCAN 14) and focused on the tool steel surface with a focal length of f = 163 mm. The focus diameter (1/e2) is about 2w0 = 63 μm (measured with a camera-based beam profiler) at a theoretical Rayleigh length of about ZR = 3030 μm. To obtain reproducible and correct average overlapping pulses, the feed per pulse (i.e., scan speed vscan) and the line distance (LD) are fixed at 14 μm as well as the pulse repetition frequency at 500 kHz.

1. Measuring equipment and target figures

The central objectives of the presented work are the investigation of productivity and quality of laser texturing and polishing,

dVdt=vscanLDNzabl.
(1)

Productivity in the context of laser texturing is quantified by the volume ablation rate dV/dt (mm3/min), which can be determined as shown above depending on the scan speed vscan, the line distance LD, the ablation depth zabl, and the number of repeats N. Productivity in the context of laser polishing is defined as the surface rate SRN (cm2/min) without the dimension in the z-direction and can be calculated as follows:

SRN=vscanLDN.
(2)

Nonproductive idle times such as deceleration and acceleration of the galvanometer scanner are not taken into account for process development. In the evaluation of the overall process for surface texturing, the whole processing time is included (cf. section “System technology” and section “Impact on the overall process”). The ablation rate is averaged over the entire ablation process. In addition, the power specific volume ablation rate dV/dt/Pav for measuring the efficiency ϵA of the ablation process is determined as an auxiliary quantity. The quality of laser machined surfaces is quantified by the surface roughness Sa (μm). The Sa value is calculated from a surface topography measurement regarding DIN EN ISO 2517819 and can be seen as an extension of the most common Ra value in one dimension,

Sa=1AA|z(x,y)|dxdy.
(3)

The depth of ablation and the roughness are the two central measured variables of the experimental investigation. They are both determined using the Keyence VK-9700 Laser Scanning Microscope, which uses a semiconductor laser with a wavelength of λL=408nm for the topography measurement. The vertical resolution is about 0.5 μm for measuring the depth and about 0.1 μm for measuring the roughness. The lateral resolution of the Sa images is about 0.43 μm. The raw data are further processed by the software Mountainsmap from Digital Surf. The determination of the ablation depth zabl takes place via automatic step recognition, in which the height difference of two averaged planes is calculated. By using a phase-correct Gaussian profile filter with different cut-offs, a wavelength dependent Sa spectrum can be achieved as well as a global Sa value without filtering.

2. Surface texturing

Fluence is known as the dominant influence on productivity and quality for ultrashort pulsed laser ablation. Since the central goal of the work is the increase of productivity, the fluence F0 is varied beyond the efficiency maximum of the single pulse ablation F0,Optimum=e2Fthr20–22 up to a maximum power of 312 W. In these kinds of process regimes, the burst mode is able to transfer more energy to the target material than single pulses due to the possibility to generate molten surfaces with high quality where single pulses would lead to crater formation destroying the surface quality. As a result of this, the variation of the number of pulses per burst (PpB) is of great interest besides the fluence. The energy expended increases linearly with the number of pulses per burst while the single pulse peak fluence remains constant. That means that a burst of ten pulses contains ten times more energy than a single pulse. For reasons of benchmark, surface texturing was also done by a 10 ps ultrashort pulse laser source from the company Edgewave as well as by an industry established 400 ns laser source of IPG.

In Fig. 5, the pulse duration for various single pulse peak fluences F0 as a function of the power specific volume ablation rate dV/dt/Pav defined as process efficiency ϵA is shown. The experimental data presented show different pulse durations of 2 and 10 ps as well as for 400 ns for comparison purposes. It can be observed that the ablation efficiency increases by a factor of more than two for each applied fluence by reducing the pulse duration from 10 to 2 ps. The maximum efficiency for 10 ps amounts to 0.08 mm3/min/W while the efficiency for 2 ps exceeds 0.18 mm3/min/W. The ablation efficiency for ns-processing is in the range between 0.13 and 0.18 mm3/min/W. But the used fluence is at least one order of magnitude higher than for ultrashort pulse processing. However, for benchmark reasons, the range is plotted as well in the same diagram. These results observe that an ultrashort pulse laser source is as efficient as a nanosecond laser source—by applying pulses in the range of 2 ps. For single pulse ablation, the effect is already observed for several materials (steel, copper, and Inconel).23–26 

FIG. 5.

Influence of the pulse duration and applied fluence on the ablation efficiency. A reduction in pulse duration in the USP regime from 10 to 2 ps leads to an efficiency increase by a factor of more than 2 so that the same efficiency is reached like for nanosecond ablation.

FIG. 5.

Influence of the pulse duration and applied fluence on the ablation efficiency. A reduction in pulse duration in the USP regime from 10 to 2 ps leads to an efficiency increase by a factor of more than 2 so that the same efficiency is reached like for nanosecond ablation.

Close modal

By using pulse bursts in the ultrashort pulse processing, the absolute ablation efficiency drops as a result of shielding effects due to the shorter pulse-to-pulse time period (approximately 25 ns). However, also for burst processing, the ablation efficiency increases by decreasing the pulse duration.27 Compared to nanosecond processing, there is also, for burst processing (2 ps), a broad range of parameter settings that is as efficient as the benchmark (cf. Fig. 6).

FIG. 6.

Applying pulse bursts, the efficiency slightly drops due to shielding caused by pulse-to-pulse interaction. Nevertheless, the efficiency is in the same range like for nanosecond.

FIG. 6.

Applying pulse bursts, the efficiency slightly drops due to shielding caused by pulse-to-pulse interaction. Nevertheless, the efficiency is in the same range like for nanosecond.

Close modal

In Fig. 7, the corresponding ablation rates depending on the applied number of pulses per burst and single pulse fluence are illustrated. By increasing the pulses per burst and the fluence, the ablation rate can be improved significantly. While single pulse processing with 1.0 J/cm2 achieves an ablation rate of 1.49 mm3/min, 10 PpB reach 10.15 mm3/min. For single pulses with a fluence of 4 J/cm2, a maximum ablation rate of 3.57 mm3/min is measured. Applying the same single pulse fluence to ten pulses within a burst, a maximum ablation rate of 42.42 mm3/min is achieved (cf. Fig. 7). This value is more than one order of magnitude higher compared to ablation rates achieved with single pulse processing. However, the applied energy is also one order of magnitude higher. Nevertheless, only the burst technology enables to transfer such an amount of energy to the surface while achieving appropriate surface qualities (cf. Fig. 8) using galvanometer scanners. By comparing the ablation rate of burst processing of a 2 ps ultrashort pulse laser source with a conventional industry used nanosecond laser source, the achieved higher ablation rates are apparent. The maximum transferred ultrashort pulse power is about 312 W in this experimental trial.

FIG. 7.

Dependency of the absolute volume ablation rate on the applied number of pulses per burst for a pulse duration of 2 ps. Partially exceeding ablation rates for ultrashort pulse ablation compared to the nanosecond benchmark. Up to 312 W of maximum transferred USP averaged power.

FIG. 7.

Dependency of the absolute volume ablation rate on the applied number of pulses per burst for a pulse duration of 2 ps. Partially exceeding ablation rates for ultrashort pulse ablation compared to the nanosecond benchmark. Up to 312 W of maximum transferred USP averaged power.

Close modal
FIG. 8.

Corresponding dependency of the surface quality on the applied number of pulses per burst for a pulse duration of 2 ps. Broad range of lower surface roughness for high power USP ablation compared to the benchmark of nanosecond. Minimum surface roughness of <0.5 μm for 5 PpB.

FIG. 8.

Corresponding dependency of the surface quality on the applied number of pulses per burst for a pulse duration of 2 ps. Broad range of lower surface roughness for high power USP ablation compared to the benchmark of nanosecond. Minimum surface roughness of <0.5 μm for 5 PpB.

Close modal

In Fig. 8, the measured corresponding roughness Sa for the investigated PpB as a function of the fluence F0 is shown. In the burst mode, the roughness takes a maximum of Sa>10.5μm in the range of F0<1.0J/cm2. By increasing the fluence, an extreme decrease of roughness occurs. For 5–7 PpB, the roughness decrease results in a global minimum of Sa0.5μm, which lies in the fluence range of F0=1.02.0J/cm2. A further increase in the fluence leads to a successive roughening of the surface up to a roughness of about 1 μm. For 8–10 PpB, a minimum Sa0.6μm is reached for a fluence of F0=1.0J/cm2. An increase in the fluence leads to a steeper increase of the roughness than for less pulses per burst that ends in a maximum roughness of about 6 μm for 10 PpB. The highlighted range for surface roughness of nanosecond processing shows that a broad range exists where a better surface quality can be achieved using ultrashort pulses due to a slightly molten surface.

3. Laser polishing

As a result of laser ablation, the 3D structures observe a surface roughness of Sa0.5μm depending on the process parameters used. This often does not meet customer requirements, so that an additional polishing process must be carried out afterwards. Particularly with regard to the molding of transparent plastics, the most exacting demands are placed on the surface quality of the molding tool. Since the structure sizes are in the range of a few micrometers, manual polishing is out of the question for the subsequent process step. Machine-supported polishing processes also fail due to the sometimes complex and small structure sizes.

In the underlying research project, the removal of the microroughness with a spatial wavelength of λ<80μm is investigated by the use of ultrashort pulse laser sources. Conventionally, the great advantage of ultrashort pulse machining is that the laser material interaction process is dominated by evaporation. This means that due to the high pulse peak intensities, the phase transition of the material from solid to gaseous state does not observe a molten phase. However, this molten phase is needed to smooth the roughness peaks and polish the workpiece.

During burst processing, when pulses with a frequency in the megahertz range hit the material surface, the energy deposited in the material cannot be removed from the processing zone by thermal conduction before the next laser pulse hits the surface after a few nanoseconds. A local heat accumulation occurs, so that the following pulse of the same burst hits a locally preheated workpiece, whose temperature rises successively as a consequence. When the melting temperature TS is exceeded, a melting film is induced locally. The process control must be ensured that the evaporation temperature TV is not exceeded. This can be controlled by the supplied energy per pulse (EP), by the number of pulses in the burst (PpB), by the burst energy (EBurst), by the time interval of the pulse bursts (Δtrep=1/frep), and/or by the time interval of the pulses within the burst (Δtseed=1/fseed). Due to the ultrashort pulse duration and the associated high intensity, the thermals developing in the material are independent of the pulse duration. In a single pulse process, the exposure time by ultrashort pulse laser (ps-fs) is shorter than the time for melt pool generation. As a consequence of the correct temporal and spatial energy deposition, it is also possible to induce a local melt film with ultrashort pulsed laser radiation, whose remelting depth can be finely controlled due to the shortened exposure time of the ultrashort laser pulses. This process is particularly advantageous for the selective polishing of microstructures (cf. Fig. 9).

FIG. 9.

3D geometries structured with ultrashort pulses, which were subsequently selectively polished by ultrashort pulses.

FIG. 9.

3D geometries structured with ultrashort pulses, which were subsequently selectively polished by ultrashort pulses.

Close modal

As a result, the use of bursts in the ultrashort pulse range can achieve smoothing of micro roughness of three-dimensional microstructures and generate a specific gloss level. In combination with the ultrashort pulse texturing, areas with high resolution (pixel size 1–20 μm) can be selectively polished in one machine with one clamping. The 3D geometries with a structure height of about 150 μm are retained and are not remelted (cf. Fig. 9). The roughness profile shows that even a structure produced by ultrashort pulses can benefit from the polishing due to lowering the amplitudes (cf. Fig. 10). The resulting polishing rate SR is about 12.15 cm2/min.

FIG. 10.

Roughness profile of a USP polished and a USP structured sample with indication of the total surface roughness Sa.

FIG. 10.

Roughness profile of a USP polished and a USP structured sample with indication of the total surface roughness Sa.

Close modal

1. Impact of scanner algorithms

The inertia-less laser beam as a tool for materials processing can be positioned on the surface some orders of magnitude faster as compared to mechanical tools. But the dynamics is still limited by the inertia of the galvanometer scanner mirrors. To ensure a constant deposition of energy and a stable laser process, the velocity of the laser focus has to be kept in a narrow range, while the laser is interacting with the material to be processed. To this end, the galvanometer scanner controller provides a mode of operation, where the programmed vectors (processing paths) are extended by acceleration and deceleration paths. When the scanner position reaches the start of the vector to be processed, the target velocity of the laser focus has already been reached and the laser is switched on. This mode is commonly called “sky-writing.” For many applications, this is a very effective approach to ensure a high quality result of the laser process. But if the laser process is dominated by an enormous number of very small features or—in other words—the scanner path consists of many very short intermitted vectors, this simple algorithm becomes ineffective. Especially in the case, if the acceleration path of a vector overlaps with the preceding vector, the scanner will perform unnecessary back and forth movements. As the probability of this case correlates with the scanner velocity, this leads to the phenomenon that increasing the scanner velocity does not lead to the expected reduction of the total processing time.

Thus, Scanlab has developed a more sophisticated algorithm to speed-up laser surface texturing: The scanner control software examines the scanner program and decides based on a set of rules if a sequence of two vectors has to be split up to insert an acceleration path. If not, the sequence can be processed in one continuous movement. A set of tolerance parameters enables the user to fine-tune the behavior of this algorithm.

Laser texturing of 3D free-form surfaces not only requires the dynamic positioning of the laser beam but also the synchronous adaption of the focal length of the optical system to maintain the laser focused on the surface while the beam is scanned over the workpiece. In most galvanometer-based 3D laser scanner systems, a lens in a telescope upstream in the beam path is moved by a linear actuator to shift the focus. Inherently, this linear movement of the lens is less dynamic as the rotation of a scan mirror by a factor of 3–5. For 3D processing, this had been a fundamental limit for the applicable scanner velocity. To circumvent this hurdle, alternative approaches have been investigated to shift the focus of the laser beam. Scanlab's innovative excelliSHIFT technology relies on a reflective optical system that transforms an angular movement of a scan mirror to a focus shift. With this operation principle for all three axes of the 3D scanner system, the same type of actuator can be employed, and the dynamics in 3D is inherently equivalent to 2D. Another approach is investigated by the Münster University of Applied Sciences using a unimorph deformable mirror.

2. Fast focus-shifter based on a unimorph deformable mirror

State-of-the-art galvanometer scanners already provide highly dynamic and precise x-y beam deflection. However, focus-shifters (“z-shifters”) relying on conventional optics are restricted to a bandwidth of a few hundred hertz. A deformable mirror placed in the path of the laser beam could be used to adjust the focal position. An efficient process rate shall be maintained; hence, the actuation rate of the deformable mirror must take place on a time scale of submilliseconds. The mirror must have a stroke of several micrometers over an aperture of a few millimeters. In the frame of this project, a fast z-shifting mirror with diffraction-limited surface fidelity was developed, which allows for an actuation rate of 2 kHz.

The mirror is based on the unimorph mirror technology, which has been developed in the past at the Photonics Laboratory for the correction of aberrations in high power lasers.28 This kind of mirror provides a promising approach to achieve high dynamic bandwidth, which is made possible by its small mass paired with high mechanical stiffness. The mirror is made from a structured piezoelectric disc, which is bonded to a highly reflective, super-polished N-BK10 glass disc with a surface roughness of <1.0 Å RMS (cf. Fig. 11). The dielectric HR-coating on the front side has a reflectivity of >99.998% over the active optical aperture of 14 mm ensuring high power handling capability. The backside electrode of the piezo element is subdivided into a large central electrode and eight keystone-shaped electrodes forming an outer ring outside the active optical aperture. The ring electrodes allow for dynamic correction of the elliptical beam footprint under oblique incidence of the laser beam on the workpiece. The mirror is actuated in a closed-loop control. The feedback signal is provided from a chromatic-confocal measurement system CHRocodile 2 SE from Precitec (see Fig. 1, top right).

FIG. 11.

Top/middle: Schematic view of the z-shifter. Bottom: photos of the manufactured mirror prototype.

FIG. 11.

Top/middle: Schematic view of the z-shifter. Bottom: photos of the manufactured mirror prototype.

Close modal

Figure 12 shows interferograms as well as the corresponding Zernike expansion of the actuated mirror surface. The surface was measured with a high-resolution phase-shifting-interferometer and fitted using 99 Zernike polynomials. The presented coefficients are given in the Wyant notation.29 In the current configuration, the mirror is able to generate a focus shift of Δz > 60 mm by using a 250 mm F-theta-lens. Possible advantages versus the innovative excelliSHIFT are the increased dynamics, a smaller footprint due to the compact design, and thus a possible integration in the scanner head instead of an installation upstream in the beam path.

FIG. 12.

Top: Interferograms of the actuated mirror surface for increasing defocus target values evaluated over an aperture of 14 mm. Bottom: Corresponding Zernike coefficients.

FIG. 12.

Top: Interferograms of the actuated mirror surface for increasing defocus target values evaluated over an aperture of 14 mm. Bottom: Corresponding Zernike coefficients.

Close modal

For every layer of the ablation process, the 3D input model is divided into patches consisting of disjoint triangle sets, so that for each patch, angle of incidence and axial movement are minimal when processed. The laser process is guided by a 2D gray scale image—a height map—that is mapped to the surface of the input model. For every layer sampled, pixel values indicate whether a tool path segment has to be created. Using the texture coordinates of a triangle, the position in the 2D domain can be determined. In a first step, tentative parallel tool paths on the entire 2D image domain are created and the segments that are covered by triangles are kept. From those segments, all parts that should not be processed in the current layer are removed according to the height map. Finally, the segments are mapped back to the 3D model and connected by travel ways, usually in a meandering pattern. However, in cases where adjacent triangles are disconnected in the UV layout—e.g., if the layout was cut open to reduce texture mapping distortion—tool paths can misalign over UV seams. This can be fixed by requiring the shifts at UV seams to be integer in the u- and v-direction, respectively. In situations where the same (tileable) height map is seamlessly repeated several times on the mesh surface, this is already the case.

However, in this approach, two types of artifacts can occur. Firstly, processing all layers in the same direction reveals the scanning pattern. Secondly, as patches are rendered independently, the triangle edges between two neighbor patches are emphasized by the laser stopping in one patch and restarting at the other one at the same position. The first problem can be addressed by rotating every layer by a random amount. To be able to still guarantee all paths align correctly, the rotation angle is approximated such that any shift along integer coordinates does not change the path layout. Simply put, this is the case when the paths in the UV layout are tileable [compare Figs. 13(b) and 13(c)]. Additionally, for all rotations except for multiples of 90°, the spacing has to be corrected [Figs. 13(d) and 13(e)]. For simplicity, the path spacing in the example is quite low. Typically, the number of path lines is much higher and corresponds to the resolution of the height map.

FIG. 13.

Impact of rotation α and spacing s on path continuity. (a) Tool paths with α = 0° and s = 0.25. (b) With α = 34° and s = 0.25, paths are discontinuous. (c) Approximating α by atan(¾) ≈ 36.87° makes the UV layout tileable again for the initial choice of s. (d) The larger α (here: 45°), the more the spacing deviates from the initial value s (here: 0.177). (e) Within certain bounds, the spacing can be corrected. Here, α = 45° and s = 0.236.

FIG. 13.

Impact of rotation α and spacing s on path continuity. (a) Tool paths with α = 0° and s = 0.25. (b) With α = 34° and s = 0.25, paths are discontinuous. (c) Approximating α by atan(¾) ≈ 36.87° makes the UV layout tileable again for the initial choice of s. (d) The larger α (here: 45°), the more the spacing deviates from the initial value s (here: 0.177). (e) Within certain bounds, the spacing can be corrected. Here, α = 45° and s = 0.236.

Close modal

In order to conceal patch boundaries, all tool paths were shifted by a small amount. For that a simple function was used that yields the same output for path segments with potentially shared endpoints. To guarantee tileable results, we use the fractional part of v as input to the shift function. The maximum offset is a user-controlled parameter. In the example in Fig. 14(c), the offset magnitude is equal to the path spacing; in Fig. 14(d), spacing is 20 μm while the maximum offset is 200 μm. Path segments that end at a mesh boundary are not shifted.

FIG. 14.

(a) Rotating paths lead to disconnections at UV seams. (b) Approximating rotation angles by the arctangent of a carefully chosen integer ratio, connection of path segments at UV seams can be restored. In this case, there is a 180° rotation over the UV seam. (c) Tool paths are shifted to prevent emphasis of patch boundaries on the final work piece. Different colors denote different patches. (d) Excerpt of tool paths from the first layer of an ablation process. The tiled pyramidal height map is rendered in the background. (e) Tool paths (incl. travel ways) for layer 6 (of 40).

FIG. 14.

(a) Rotating paths lead to disconnections at UV seams. (b) Approximating rotation angles by the arctangent of a carefully chosen integer ratio, connection of path segments at UV seams can be restored. In this case, there is a 180° rotation over the UV seam. (c) Tool paths are shifted to prevent emphasis of patch boundaries on the final work piece. Different colors denote different patches. (d) Excerpt of tool paths from the first layer of an ablation process. The tiled pyramidal height map is rendered in the background. (e) Tool paths (incl. travel ways) for layer 6 (of 40).

Close modal

The principle of online process monitoring is based on the observation of significant indicators describing the properties of the workpiece in the interaction zone and its adjacent area. The technology for acquisition and processing these indicators is useful only if it can respond clearly to a significant change of the process conditions or the resulting quality. Process control systems must operate in a contactless manner, so that they exert no adverse influence on the interaction zone. In laser material processing, this requirement is not a problem, because the processes are accompanied by several effects that can be easily and reliably observed from a distance.

OCT technology is an imaging technique based on low-coherence interferometry (LCI). It is a long-established medical examination procedure. An interferometer with a light source of low coherence length is used to measure distances and the composition of human tissue, e.g., the cornea. The short coherence length is achieved using light sources that emit light with a broad spectrum. The applied light sources are typically super luminescent diodes with a range of some 10 nm or a Swept Source Laser. In 2006, Precitec Optronik GmbH launched a thickness and distance sensor based on spectral domain OCT, and this was adapted to material processing applications with laser sources of high beam quality. The adapted technology allowed distance measurement to the required accuracy of about 10 μm, even over long distances.

However, the real innovation and, thus, the basis of a technological leap in the field of process monitoring/control is the fact that the accuracy of the interferometric measurement is not affected by the electromagnetic emissions emerging from the processed area due to beam-material interaction. The intensely bright emissions caused by the high power beam-material interaction are not coherent with the light emitted by the low coherent light source of the measuring system, and thus, only the light from the measurement system is involved in interference between the reference and the measuring path. Based on an accurate adjustment of the measurement beam coaxial to the processing laser, this technology for the first time provided an exact measurement of the depth of the keyhole in cw laser processing or the amount of ablated material when using short or ultrashort pulsed laser sources, independent of seam geometry or the processed material. The only restriction is in the dimension of the measurement point compared to the spot size of the processing beam and the measuring range in the axial direction.

In the joint project “eVerest,” the group of research institutes, component manufacturers, machine builders, and end users first time investigated the use of the low-coherence interferometry as sensor technology for laser microprocesses as a quality control device. The need was to measure the ablation depth in situ and control the process. With the 70 kHz measuring frequency of the sensor device, it was possible to achieve this goal. One of the results of the work is that in the majority of applications an ultrashort pulsed laser with scanner technology is used to deflect the laser beam, and the sensor must be adapted to this particular processing head. This is especially true with respect to the light source used for the interferometric sensor. In this case, the light source of the OCT system was selected carefully in order to reduce the chromatic shift induced by the F-Theta lens in the scanner device. Based on the findings gained in “eVerest,” this technology is able to exactly measure topography structures in the claimed accuracy (cf. Fig. 15); nevertheless, the speed of the measurement has to be adapted to the repetition rate of typical ultrashort pulsed laser devices. Based on the state-of-the art today, the only OCT technology able to match the speed of typical ultrashort pulse laser devices is the swept-source OCT (SS-OCT). Typical Fourier-Domain OCT (FD-OCT) systems today perform at a maximum acquisition rate of 250 kHz;30 for SS-OCT systems, the highest speed available today is 1.6 MHz.31 

FIG. 15.

Topography measurement by the Precitec OCT system of the treated workpiece.

FIG. 15.

Topography measurement by the Precitec OCT system of the treated workpiece.

Close modal

The individual components were incorporated into a machine based on the Lasertec 125 from DMG Mori. With the integration of the innovative excelliSHIFT of Scanlab, the dynamics known from 2D could be transferred to 3D processing. In combination with the intelligent preprocessing of the scanner job, the throughput for laser texturing could be tripled. The surface texturing of a large-format 3D mold tool for an instrument panel of a VW Up, for example, usually does take up to 3 weeks. Finally, the structuring and selective polishing do only take 7.5 days (cf. Fig. 16).

FIG. 16.

(a) 3D mold tool for an instrument panel of a VW Up clamped in a DMG LT125 laser texturing machine. (b) Impression of the applied process of laser polishing.

FIG. 16.

(a) 3D mold tool for an instrument panel of a VW Up clamped in a DMG LT125 laser texturing machine. (b) Impression of the applied process of laser polishing.

Close modal

Laser polishing provides the possibility to highlight special areas of the surface texture. This process step in the end adds a special value to the product. To do so by using ultrashort pulses is an innovation that is developed by the Fraunhofer ILT.

Processing with high power ultrashort pulse laser sources additionally gives the opportunity to gain the same throughput like with nanosecond. Nevertheless, after ablation, a black oxide layer remains on the surface. Finally, a photonic process chain is established consisting of USP ablation, USP cleaning, and USP polishing in one machine with one clamping. This reduces reworking in a massive way. Figure 17 shows the strategy of a sequentially photonic process chain to establish a high productive mold tool texturing on an excellent quality level.

FIG. 17.

Strategy of applying a photonic process chain consisting of the steps: USP ablation, USP cleaning, and USP polishing with process images and final results.

FIG. 17.

Strategy of applying a photonic process chain consisting of the steps: USP ablation, USP cleaning, and USP polishing with process images and final results.

Close modal

A further impression and final application is shown in Fig. 18, where a steering wheel cover is selectively polished after laser texturing. The corresponding final plastic part does show the glossy parts as well:

FIG. 18.

(a) Laser textured mold of a steering wheel cover with selectively polished parts. (b) The final injection mold plastic part shows the polished highlights as well.

FIG. 18.

(a) Laser textured mold of a steering wheel cover with selectively polished parts. (b) The final injection mold plastic part shows the polished highlights as well.

Close modal

Laser surface texturing is becoming increasingly important in the industry for tool and mold making. Although nanosecond based laser texturing is still established in the industry, the processing time is still a challenge. Also the ultrashort pulse laser still have shown their advantages regarding highly precise ablation in diverse applications—however, the general opinion remains that ultrashort pulse lasers are less productive.

In the collaborative project “eVerest,” a consortium of five industrial partners and three research institutes were working together on improving the processing speed of the conventional industrial established process as well as closing the gap for a broad utilization of ultrashort pulse laser sources in an industrial environment for surface texturing. To optimize the processes, the efficiency of all the components was examined. A high power ultrashort pulse laser source was developed with an output power of several 100 W. The innovative approach here is the realization of a fiber coupling unit with a transmission beyond 90% at a power level of 135 W. The improvement of the system technologies in the context of an intelligent and fast beam deflection observed a significant increase in the throughput by a factor of 3. Applying these improvements, a mold of an instrument panel can now be laser textured in just 7.5 days after 3 weeks. The technology of a unimorph deformable mirror is promising for further improvements of a faster beam deflection in the z-direction.

A detailed understanding of the process itself is essential whatever the application. The key is to combine this with modifications to the process technology and comprehensive control software. The investigations in ultrashort pulse laser texturing and polishing reveal the ultrashort pulse laser source as an universal tool. On the one hand side, processes can be as productive as with nanosecond pulses while applying pulses with a pulse duration in the range of 2 ps together with pulse bursts. On the other hand, ultrashort pulses can be used for cleaning and polishing within the same machine with one clamping. With this kind of photonic process, chain rework is reduced significantly. The aim was to make the processes much faster while simultaneously achieving even higher texturing quality. With this approach, a volume ablation rate of approximately 10 mm3/min can be achieved with a surface roughness of Sa=0.55μm by the use of an ultrashort pulse laser source.

The offline data planning and tool path generation in advance are crucial aspects. In order to conceal patch boundaries, for example, a shift of a small amount was implemented for all tool paths. The OCT technology finally enables an in situ process control for measuring the ablation depth. Based on the findings gained in “eVerest,” this technology is able to exactly measure topography structures in the claimed accuracy.

The collaborative research project “eVerest” is being carried out on behalf of the Federal Ministry of Education and Research BMBF under the funding code 02P14A145.

1.
G.
Ryk
and
I.
Etsion
, “
Testing piston rings with partial laser surface texturing for friction reduction
,”
Wear
261
,
792
796
(
2006
).
2.
J.
Schneider
,
D.
Braun
, and
C.
Greiner
, “
Laser textured surfaces for mixed lubrication: Influence of aspect ratio, textured area and dimple arrangement
,”
Lubricants
5
,
1
12 (
2017
).
3.
A.
Dunn
,
K. L.
Wlodarczyk
,
J. V.
Carstensen
,
E. B.
Hansen
,
J.
Gabzdyl
,
P. M.
Harrison
,
J. D.
Shephard
, and
D. P.
Hand
, “
Laser surface texturing for high friction contacts
,”
Appl. Surf. Sci.
357
,
2313
2319
(
2015
).
4.
Frost & Sullivan: Automotive Interior Surface Materials Market in North America—Insights and Trends, 9AB7-39 (
2016
).
5.
H.-W.
Wiederoder
, “
Fotochemisches ôzen von formgebenden Werkzeugen
,”
Swiss Mater.
9
,
9
17
(
1997
).
6.
M.
Köhler
,
Ätzverfahren für die Mikrotechnik
(
Wiley-VCH
,
Weinheim
,
1998
), p.
90
.
7.
A.
Piqué
,
R. C. Y.
Auyeung
,
H.
Kim
,
N. A.
Charipar
, and
S. A.
Mathews
, “
Laser 3D micro-manufacturing
,”
J. Phys. D Appl. Phys.
49
,
1
24
(
2016
).
8.
M.
Henry
,
P. M.
Harrison
,
I.
Henderson
, and
M. F.
Brownell
, “
Laser milling: A practical industrial solution for machining a wide variety of materials
,”
SPIE Proc.
5662
,
627
632
(
2004
).
9.
M.
Kuhl
, “
From macro to micro–The development of laser ablation
,” in
Congress Proceedings ICALEO 2002
, edited by E. Beyer (LIA, Orlando, FL, 2002), Vol. 94, p.
168615
.
10.
B. N.
Chichkov
,
C.
Momma
,
S.
Nolte
,
F.
Alvensleben
, and
A.
Tünnermann
, “
Femtosecond, picosecond and nanosecond laser ablation of solids
,”
Appl. Phys. A
63
,
109
115
(
1996
).
11.
Y.
Zhu
,
J.
Fu
,
C.
Zheng
, and
Z.
Ji
, “
Effect of nanosecond pulse laser ablation on the surface morphology of Zr-based metallic glass
,”
Opt. Laser Technol.
83
,
21
27
(
2016
).
12.
K.-H.
Leitz
,
B.
Redlingshöfer
,
Y.
Reg
,
A.
Otto
, and
M.
Schmidt
, “
Metal ablation with short and ultrashort laser pulses
,”
Phys. Proc.
12
,
230
238
(
2011
).
13.
A.
Dohrn
and
A.
Gillner
, Application Center. See https://www.ilt.fraunhofer.de/content/dam/ilt/en/documents/product_and_services/laser_material_processing/B_Application_Center_Laser_Structuring_for_Tool_and_Mold_Construction_2015.pdf for “Laser Structuring for Tool and Mold Construction.” Retrieved 18 December 2019.
14.
See https://amplitude-laser.com/produit/tangor/ for “Amplitude: Tangor Product Sheet.” Retrieved 18 December 2019.
16.
See https://www.edge-wave.de/web/produkte/ultra-short-pulse-systeme/fx-serie/ for “Edgewave: Ultrashort Pulse InnoSlab Lasers. FX-Series.” Retrieved 18 December 2019.
17.
See https://www.amphos.de/products/ for “Amphos: High Power Ultrafast Lasers.” Retrieved 18 December 2019.
18.
A.
Brenner
,
B.
Bornschlegel
, and
J.
Finger
, “
Increasing productivity of ultrashort pulsed laser ablation for commercialization of micro structuring
,” in
Nanoengineering: Fabrication, Properties, Optics, and Devices XV
, edited by A. E. Sakdinawat, A.-J. Attias, B. Panchapakesan, and E. A. Dobisz (SPIE,
2018
), Vol. 10730, p.
107300H-1-12
.
19.
Deutsches Institut für Normung
,
DIN EN ISO 25178-2. Geometrische Produktspezifikation (GPS)—Oberflächenbeschaffenheit: Flächenhaft—Teil 2: Begriffe und Oberflächen-Kenngrößen (25178-2)
(
Beuth
,
Berlin
,
2012
).
20.
B.
Neuenschwander
,
G. F.
Bucher
,
C.
Nussbaum
,
B.
Joss
,
M.
Muralt
,
U. W.
Hunziker
, and
P.
Schuetz
, “Processing of metals and dielectric materials with ps-laserpulses. Results, strategies, limitations and needs,” in
Laser Applications in Microelectronic and Optoelectronic Manufacturing XV
, edited by H. Niino, M. Meunier, B. Gu, and G. Hennig (SPIE, 2010), p.
75840R
.
21.
B.
Lauer
,
B.
Jäggi
, and
B.
Neuenschwander
, “
Influence of the pulse duration onto the material removal rate and machining quality for different types of steel
,”
Phys. Proc.
56
,
963
972
(
2014
).
22.
B.
Neuenschwander
,
B.
Jaeggi
,
M.
Schmid
, and
G.
Hennig
, “
Surface structuring with ultra-short laser pulses: Basics, limitations and needs for high throughput
,”
Phys. Proc.
56
,
1047
1058
(
2014
).
23.
B.
Jaeggi
,
B.
Neuenschwander
,
M.
Schmid
,
M.
Muralt
,
J.
Zuercher
, and
U.
Hunziker
, “
Influence of the pulse duration in the ps-regime on the ablation efficiency of metals
,”
Phys. Proc.
12
,
164
171
(
2011
).
24.
J.-T.
Finger
,
Puls-zu-Puls-Wechselwirkungen beim Ultrakurzpuls-Laserabtrag mit hohen Repetitionsraten
(
Apprimus Verlag
,
Aachen
,
2017
), pp.
86
87
.
25.
B.
Neuenschwander
,
B.
Jaeggi
,
M.
Schmid
,
V.
Rouffiange
, and
P.-E.
Martin
, “
Optimization of the volume ablation rate for metals at different laser pulse-durations from ps to fs
,”
Proc. SPIE
8243
,
824307
(
2012
).
26.
B.
Jaeggi
,
B.
Neuenschwander
,
S.
Remund
, and
T.
Kramer
, “
Influence of the pulse duration and the experimental approach onto the specific removal rate for ultra-short pulses
,”
Proc. SPIE
10091
,
100910J
(
2017
).
27.
A.
Brenner
,
B.
Bornschlegel
, and
J.
Finger
, “
Increasing productivity of ultrashort pulsed laser ablation in advance for a combination process with ns-laser
,”
JLMN
14
,
100
107
(
2019
).
28.
S.
Verpoort
,
P.
Rausch
, and
U.
Wittrock
, “
Characterization of a miniaturized unimorph deformable mirror for high power CW-solid state lasers
,” in
MEMS Adaptive Optics VI
, edited by S. S. Olivier, T. G. Bifano, and J. Kubby (SPIE,
2012
), Vol. 8253, p.
825309-01-12
.
29.
E. P.
Goodwin
and
J. C.
Wyant
,
Field Guide to Interferometric Optical Testing
(
SPIE
,
Bellingham
,
WA
,
2006
), p.
100
.
30.
See https://wasatchphotonics.com/wp-content/uploads/WP-PS_CobraS-800_10Jun19-web.pdf for “Wasatch Photonics: Cobra-S 800 OCT Spectrometer Series.” Retrieved 10 January 2020.
31.
See https://www.optores.com/images/datasheets/P40_70_20_NG-OMES_Data_Sheet_v1.0.pdf for “OptoRes: OMES 4D MHz-OCT System.” Retrieved 10 January 2020.