Pyrometer control for laser welding of transparent polymers
Johannes Eckstädta , Alexander Frankea
aLEISTER Technologies AG, Galileostrasse 10, 6056 Kaegiswil, Switzerland
Transmission welding of transparent to absorbing plastics is a proven technique for applications in various industrial branches.
It stands out due to high precision with low thermal effect on the base material of the components. Furthermore the laser is not in direct contact with the samples and has no mechanical impact on them [1-3].
Lately, the joining of both transparent components has attracted increased interest, in particular in medical engineering. The welding of transparent parts while using lasers in a range of about 1 µm wavelength requires special infrared absorbers (e. g. Clearweld by Gentex Technology Corp.) [4-5]. The additional material is expensive and extensive to apply.
With the use of a laser wavelength of about 2 µm the absorption increases to a level where welding is possible without any absorbent (Figure 1) .
Figure 1: Spectrum analysis of 2 mm PET
The importance of a reliable process control increases with the necessity of a high quality welding. Especially in the medical market a pyrometer is often used to measure or even control the temperature of the weld.
The quality of the measurement depends on the materials used, the wavelength of the laser source and the wavelength range of the detector. Therefore a new approach for an unsophisticated temperature measurement has been developed.
2. Background pyrometer
The temperature is determined by the heat radiation during welding. The Planck's spectrum of radiation is a reference model that shows the emitted wavelength spectrum of an ideal black-body with its electromagnetic energy distribution for defined temperatures (see Figure 2).
A gray body like the later used welding samples emits only a fraction of the radiation of an ideal black body depending on the material and its surface. The heat radiation can be measured by a photodiode which is the basic component of a pyrometer. It converts the incoming energy of the heat radiation into a proportional photoelectric current.
Figure 2: Planck's spectrum of radiation
The typical temperature range of polymer welding is between 400 K and 600 K. So, the emitted heat radiation starts at a wavelength of about 1 µm. The upper limit of the measurement range is about 2.5 µm because higher wavelengths are absorbed by the upper polymer of the welding and do not reach the photodiode of the pyrometer. This small range does not even include the maximum of the energetic distribution and may cause a degraded signal.
When welding at a wavelength of 2 µm the reflected and scattered laser radiation overlaps with the measured heat radiation in its wavelength and therefore falsifies the measurement of the pyrometer. Caused by its high energy level it outshines the heat and makes the measurement useless. So it is obligated to separate laser radiation from heat radiation. An optical filter with a special coating has been developed to solve this problem. This so called notch filter has a high transmission (almost 100 %) in a wavelength range of 1 µm to 2.5 µm with a small excluded sector or “notch” at a wavelength of the laser where the transmission is almost 0 %. In this case the “notch” is around 1960 nm ± 50 nm (Figure 3). The filtered wavelength range has to be small, so that the signals measurability does not fall off in quality.
Figure 3: Characteristics of the notch filter
3. Control loop for welding
Besides simple monitoring, the pyrometer is developed for use in an active control loop.
The target temperature is controlled by switching between two different laser powers. If the measured temperature is below the target temperature the upper laser power is set to increase the temperature – for higher temperatures it is switched to the lower power.
In welding trials of PET samples a laser beam is guided along a rectangular shape with different radii in the corners (Figure 4).
Figure 4: Beam guidance over the PET samples (dimensions in mm)
The laser optic is moved by an axis system along the welding path. In order to counteract inertia with even motion, axes slow down speed in curves.
So the feed of the laser beam varies along the welding path. When the laser power is constant, the energy input increases with lower feed and may cause local burning in the corner. These deviations in the welding seam shall be compensated by the pyrometer control.
For examination of the benefit of the pyrometer control two 2 mm thick PET samples are welded along the path showed in Figure 4. After basic studies for suitable parameters the welding process is going to be monitored by the pyrometer without control loop. Afterwards the PET samples are welded with same parameters in active control mode. For good comparability the target temperature should be near to the temperature measured in the monitoring run.
An exact calibration of the photodiode current to a temperature is related to the transmissivity of the materials used in the optical path of the pyrometer. To remain generally valid pyrometers are calibrated to a black body. So, an optimized calibrated signal of a welding zone of PET differs to the one of a black body. For this reason the tests have been performed using abstract values instead of temperature.
Without the notch filter the pyrometer measures mainly the intensity of the laser radiation instead of heat. So the pyrometer signal is always at maximum when the laser is turned on no matter what temperature prevails in the welding zone. In contrast, with the integrated notch filter the laser radiation reaching the photodiode is reduced to a minimum and the temperatures that are relevant for welding can now be monitored.
Another option is running the laser in quasi-continuous wave instead of normal continuous wave mode. In quasi-continuous wave mode the laser power is periodically turned off but is constant within these periods. In the laser-off periods a pyrometer could measure the heat radiation without the use of a notch filter. But the off-periods lead to a longer cycle time which is critical for many applications. So the costs of the notch filter are amortized within a very short period by faster cycle times.
For demonstration purposes the parameters evaluated are set to the upper limit of the process window. In that way a higher energy input can easily be spotted by the formation of bubbles in the seam.
Figure 5: Samples and pyrometer signal with and without control loop
The samples in Figure 5 are both welded with the same parameters but with and without control loop. The radii are marked in the diagram (blue dashed line). The yellow curve represents the pyrometer signal of the uncontrolled welding process. It shows that over the whole process the temperature increases continuously caused by steady heating of the material. A deflection of the temperature when passing the corners is only visible at the two most sharp corners (R=0 and R=1). With a radius of 2 mm the curve only shows an increased incline. The 4 mm radius can already be passed without slowing down so there is no change in the energy input.
The increasing temperature affects the welding as shown in the up left picture in Figure 5. The welding starts at the 9-o’clock-position and runs clockwise (Figure 4). The first half of the welding seam looks clean and consistent. But with the sharp corner (R=0) the temperature crosses the upper limit of the process window and bubbles are formed.
Deviated from the curve without control loop, the target pyrometer value is set. It is a little below the first curve so the temperature can be regulated to this constant value almost right from the start. If the pyrometer curve exceeds this value the laser power is switched to the lower power. In Figure 5 the red curve respectively the down left picture represent the welding with control loop. The pyrometer signal and with that the temperature is very steady. The first two rounded corners can only be recognized by minimal amplitudes in the pyrometer signal which does not affect the welding, yet. In the sharp corner the control loop is too slow to compensate the slowdown on time and the signal deflects - but with a minor magnitude than the uncontrolled signal. The maximum value is still under the limit where bubbles are starting to develop. With the control loop the target value is reinstated by half a second. According to the pyrometer signal the welding seam is also more uniform and does not show any bubbles. The control loop is equalizing the energy input and proves its feasibility.
A laser source using 1960 nm wavelength is suitable to weld two transparent plastic materials to each other. Monitoring and even controlling the welding process is possible using a pyrometer with a notch filter which fades out a small wavelength range around the 1960 nm. Without such a filter the prime signal detected comes from the laser source instead of the heat.
The feasibility of a closed loop control of the welding temperature has been proven in extensive tests. In comparison to welds without control loop the temperature does not rise, but stays stable. This improves the welding quality and assures a more stable process of laser welding transparent materials.
 Russek, U. A.: Prozesstechnische Aspekte des Laserdurchstrahlschweißens von Thermoplasten, Dissertation, RWTH Aachen University, Shaker Verlag: Aachen (2006).
 Grewell, D.; Benatar, A.; Park, J.: Plastics and Composites Welding Handbook, Carl Hanser Verlag, Munich, S. 271–311 (2003).
 Rösner, A.; Olowinsky, A.: Laser welding of polymers, Industrial Laser Solutions, 28(1), S. 29-31 (2013).
 Wenzlau, C.: Component Welding for Blood Analysis Device, Laser Technik Journal, S.44-46 (2014/01)
 N.N.: Clearweld® Weldable Resins Guide. URL: http://www.clearweld.com/cms-assets/documents/31194-492207.resintechbulletin.pdf, 02.08.2018
 Mingareev, I.; Weirauch, F.; Olowinski, A.; Shah, L.; Kadwani, P.; Richardson, M.: Welding of polymers using a 2 µm thulium fiber laser. Optics & LaserTechnology, 44, S. 2095–2099 (2012)