Frederik Maiwald and Stefan Hierl determine the optimal parameters for welding transparent polymers used in optical and medical devices
Optical and medical devices made of polymers are gaining popularity due to their cost advantage compared to glass. However, the typical usage of coloured absorbent additives – used to improve the absorption of laser radiation in polymers – is not sought after when welding such devices. This is because most optical and medical applications instead require purely transparent materials.
Laser transmission welding of polymers has previously been demonstrated without absorbers. However, the process is still known to lack stability and productivity. My colleagues and I therefore set out to determine the optimal parameters for achieving fast and reliable transmission welding in transparent polymers.
Process principle
Laser transmission welding is a well-known joining technology for thermoplastics. It can provide precise, reliable and hermetic sealing without particle formation or the requirement of adhesives. These advantages are essential for medical and optical applications. The welding process is used in two variants: transparent-absorbent welding and transparent-transparent welding, as seen in figure 1.
Figure 1: (a) Transparent-absorbent welding of a 70 x 85 x 130mm³ surge tank and (b) absorber-free transparent-transparent welding of a 50 x 50 x 2mm³ exemplary fluidic component.
Transparent-absorbent welding, in which the lower joining partner is coloured, has been commonly established for many years. It is frequently used in the automotive industry, for example in the manufacture of sensor housings and surge tanks. In contrast, no absorbers are required for new transparent-transparent welding processes, which are enabling components such as fluidic devices or optical sensors made of identical, clear material to be joined without colouring or other additives – while still benefiting from the above advantages.
Due to the laser-absorbing blackening of the lower joining partner in transparent-absorbent welding, the zone of laser radiation absorption and fusing is located exactly at the interface between the two partners.
When welding transparent polymers without absorbers however, two challenges emerge: the deposition of laser energy in general; and achieving the required selective energy deposition in the joining zone. To attain absorption without additives, lasers emitting in the polymers’ intrinsic absorption spectrum – between 1.6µm and 2µm – are used. However, welding with standard optics leads to poor focusing and insufficient energy concentration in the joining zone, resulting in a melting of the upper joining partner, including its upper surface (shown in figure 2a).
Figure 2: Schematic sketch (top) and images of microtome sections (bottom) demonstrating the weld seam formation in dependence on the Rayleigh length zR: Bulged surface because of poor focusing (a) and desired weld seam without surface defect (b). Material: Polyamide 6, laser power = 60W (cw), feed rate = 200mm/s.
On the contrary, if the beam is focused within the workpiece with a high numerical aperture (NA), the joining zone can be selectively melted. The Rayleigh length zR – the distance from the laser beam waist to the point where the area of the cross-section is doubled – must also be short. In doing this, the decrease of the laser beam’s cross-sectional area overcompensates the loss of laser power caused by absorption, maximising the radiation intensity in the joining zone (shown in figure 2b).
Equipment setup
Figure 3 shows our experimental setup for absorber-free laser transmission welding. A thulium fibre laser with a wavelength of 1,940nm is used. The fixed-focus setup consists of a breadboard, where a rail carrying the optical elements is mounted on slides. The beam is guided through an adjustable beam expander and a Galilean telescope with a high NA (0.6) focusing lens. A fine threaded spindle moves the rail, enabling the variation of the distance between the optics and the specimen. A clamping device with a conical slit hole fixes the two specimens (50 x 50 x 1mm³ each) in an overlap. A two-axis linear system moves the specimens at up to 300mm/s.
Figure 3: Sketch of the processing setup with a high NA focusing optic and off-axis pyrometer for process monitoring.
Regarding the weld seam quality, vertical expansion of the seam is especially crucial. Whereas an excessively large seam leads to a molten surface with visible and palpable irregularities, a seam of insufficient length causes leakage and inadequate strength. Therefore, to fulfil the high demands on weld seam quality in the medical industry, online process monitoring is necessary. A customised pyrometer based on Micro-Epsilon’s CTM-3CF1-22, with an optical filter blocking the laser wavelength, is therefore integrated off-axis into the experimental setup. Due to the filter, the sensitivity is reduced to the range of 2-2.5µm. The analogue output is processed with over 50kHz using a cRIO-9035 controller.
Welding tests
In our welding tests using the crystal clear cyclic olefin copolymer Topas 8007-04, at least seven welds per setting were processed at 60W using a 200mm/s feed rate and at five different laser focus positions (zf,rel)between 0.8mm and 1.2mm.
At zf,rel = 1.05mm, the laser focus position is in the joining zone and with increasing zf,rel, the laser focus position is shifted downwards in the material.
The values of zf,rel represent the focus shift measured in air. Due to refraction, the shift in air is less than inside the material. Figure 4a shows photographs of microtome sections of typical weld seams for different laser focus positions. The sections were photographed in polarised light using a transmission light microscope. Figure 4b schematically shows the weld seam and the measured distance between the specimen’s surface and the upper (U) and lower (L) end of the seam in dependence on the laser focus position zf,rel. Figure 4c shows the corresponding, averaged pyrometer signals, detected along a 20mm section in the middle of the specimen.
Figure 4: (a) Photographs of microtome sections of typical weld seam states and (b) distance between specimen’s surface and upper (U) and lower (L) end of the weld seam in dependence on laser focus position zf,rel. (c) Pyrometer signal averaged along 20mm seam length. Material: Topas 8007-04, laser power = 60W, feed rate = 200mm/s, Rayleigh length = 0.3mm.
The resultant seams can be divided into four classes:
Class 0: The welds of class 0 (zf,rel = 1.2mm) are inadequate as no firm connection of the joining partners is achieved. The joining partners are loose after welding, even if the visible heat-affected zone covers both partners. The temperature leading to a visible heat-affected zone is less than the melt temperature (Tmelt) needed for bonding. The pyrometer’s signal is below the threshold and only noise with a mean value of 0.8mV is detected.
Class IA,B: Class I (zf,rel= 1.1mm and 1.0mm) represents the desired result. Since line energy and focal position are well-matched, both joining partners are connected firmly and a sufficient distance U (0.1mm < U < 0.2mm) between the seam and the surface is achieved. The pyrometer signals are 1.3mV and 1.5mV, respectively.
Class II: In this marginal case (zf,rel = 0.9mm), a firm connection is achieved but the surface may or may not show noticeable defects as the weld seam just reaches the surface (U = 0). The specimens have to be checked and a focus adjustment downwards is advised. The pyrometer signal is 2.6mV.
Class III: Although both joining partners are connected tightly, the welds of class III (zf,rel = 0.8mm) are inadequate since the surface is damaged by a palpable bulge caused by the weld seam. The pyrometer signal is 4.2mV.
The experiment showed that proper and faulty parts can be differentiated using a pyrometer. Both a molten surface and a faulty joint can be distinguished from proper results. Operating the process between 1.3mV and 1.5mV signal ensures the desired result of tight seams without surface defects (marked blue in figure 4b and c). Additionally, weld seams processed with up to 2.6mV signal can be acceptable, since the surface may or may not show noticeable defects.
Conclusion and future work
Our experiments determined that by focusing the beam of a thulium fibre laser with a high NA, precise weld seams can be achieved in transparent polymers without additional absorbers. To fulfil the high demands on quality in the medical industry, the weld seam position needs to be monitored using a pyrometer, for example.
Since the signals are recorded with a frequency of over 50kHz during the process, closed-loop control of focal position or other parameters is possible. This will form part of our future work, as will the investigation of adapted intensity distributions for the process stabilisation and transformation of weld seam geometry.
Acknowledgements
Our work on process development and monitoring in laser plastic welding is supported by the European Union and the Free State of Bavaria as part of the TheCoS, 3D-LasPyrInt-Scanner and GipoWELD projects. We would like to thank the sponsors and the project partners Arges, Bayerisches Laserzentrum, Gerresheimer Regensburg, LPKF WeldingQuipment, Micro-Epsilon Messtechnik, Nexlase and the universities in Erlangen-Nuremberg and Pilsen for the good cooperation.
Frederik Maiwald and Professor Stefan Hierl work at the East Bavarian Technical University of Regensburg’s Laser Material Processing Laboratory.
This work can be read in full in Procedia CIRP, Volume 94, 2020, Pages 686-690 (https://doi.org/10.1016/j.procir.2020.09.117) and in the Journal of Laser Micro/Nanoengineering, Volume 16, Number 1, 2021 (https://doi.org/10.2961/jlmn.2021.01.2002).