Wednesday, October 3, 2018

High Precision Optics for LED Applications Made of Liquid Silicone Rubber (LSR)


Conventional light sources are steadily replaced by energy-saving and durable light emitting diodes. Their advantages over conventional light bulbs include a higher light emission with less power consumption and voltage, a longer durability and a smaller frame size [1]. Nevertheless, during operation LED still generate a significant heat generated, as only a share of about 20% of the electrical power is converted into light.
Regarding the performance of so-called high-power LED (HP-LED) the materials used in front of and behind the semiconductor chips are of major importance. Transparent liquid silicone rubber offers a promising alternative because of its good thermal properties and high UV-stability. The purpose of a research project, which was funded by the German Research Foundation (Deutsche For schung gemeinschaft, DFG), was the production of high precision optics for LED for an automotive high beam module made of liquid silicone rubber (LSR). Because of its suitability for large scale production, injection molding is used for processing of transparent LSR which is only available since 2008.
In the research project, the basic operation of the newly developed LSR-optics in a high beam module is studied. In the conventional design of a high beam module (Figure 1) the semiconductor chip of the LED is first encapsulated with a non-injection-moldable casting compound. The encapsulation is  used to protect the LED against environmental influences and mechanical stresses, represents simultaneously the first optical element and is therefore also referred to as primary optics.
In the low power range of LED commonly transparent epoxy resins or polyurethanes are used as primary optics, in high-power range due to the high temperatures and exposure to UV silicones are common [2, 3]. Because of the production in the dispensing technique, the freedom in design of these materials is restricted. As a consequence the emitted light cannot be targeted towards downstream secondary optics. Correspondingly low is the luminous efficiency of the optical system. Therefore, optical heads made of thermoplastic materials are used directly on the LED chip in order to keep a major part of the emitted light inside the optical system.
Highly transparent liquid silicone rubbers have a significantly higher temperature and UV radiation stability compared to epoxy resins, polyurethanes or thermoplastics [4]. Furthermore, they can be processed by injection molding with the associated process-specific advantages such as high freedom in design and less reworking. In order to utilize these advantages, the developed LSR-optics takes the encapsulation of the LED as well as the function of the optical head into account (Figure 1). In the conventional setting of LED light output losses up to 15% are expected at the transition between the primary optics and optical heads. By combining both elements in one component, these losses can be reduced.

Mold Concept

In order to increase the economical attractiveness, the LSR-optics is injected directly on the LED board (Figure 2). Therefore, the LED board is inserted

into the opened mold and held in position by centering pins. Afterwards, the liquid silicone rubber is injected through an opening in the back of the board with a cold runner system. Thus, cost and time-consuming mounting steps can be reduced.
To precisely determine the position of the optics on the LED board and hence, ensure a reproducible production cycle, an accurate centering of the circuit board is necessary. Therefore, the LED board is positioned with two centering pins on the nozzle side. During the injection and curing phase the sealing of the contacting area and the cooling element has to be considered. Therefore, a sealing concept was developed where mold elements are pushed on the targeted sealing edge after closing the mold. The mold has a modular design with various optical mold inserts, so that the LED types and the optical design can be varied.

Influence of the Mold Temperature and the Injection Flow Rate on the Molding Accuracy and Optical Properties

The objectives of several test series were the demonstration of the applicability of this approach and the definition of a suitable process window. The mold temperature and the injection volume flow rate were identified as crucial process parameters.
The mold temperature has a strong influence on the crosslinking behavior of the LSR. In addition, it determines the thermal stresses on the chip during injection molding. In order to exclude an unnecessarily long exposure of   the chip to heat and to design the process as economically as possible, the crosslinking time was defined depending on the mold temperature in test series: Once the optical components are demoldable, the crosslinking time is terminated. The injection flow rate on the other hand influences the shear stress on the chip during the injection phase.
Therefore, the mold temperature and the injection flow rate were varied in the process studies in a 2²-factors experimental design with central point. Thereby, the curing time was adapted in dependence of the mold temperature in order to achieve a similar curing degree at the ejection point. To quantify quality, the molding accuracy and the optical properties were measured and evaluated.
In order to analyze the molding accuracy the surface profile in the center of the LED optics and the mold insert were analyzed. The correlation between the molding accuracy and the varied process parameters was analyzed using the average and the maximum deviation between mold insert and optics (Figure 3). The average deviations in x-direction are between 2 and 6 µm. The maximum deviations, depending on the testing point, are between 12 and 24 µm. In y-direction the deviations are higher, between 6 and 10 µm in average and between 14 to 24 µm at maximum.
Figure 3. Average and maximum difference independence of the mold temperature and injection volume flow [5]

Nevertheless, the same tendencies regarding the varied processing parameters are seen for both directions. The injection flow rate has no significant influence on the molding accuracy. However, the molding accuracy decreases with higher mold temperatures. Especially higher mold temperatures can lead to a post-crosslinking, which does not take place under mold constraint. Moreover, the crosslinking time can only be adjusted roughly to the mold temperature so that an identical degree of crosslinking at the time of mold release is not guaranteed.
Independent from the varied processing parameters, the LEDs are functional after the injection molding process. In order to test the optical properties,  the maximum illuminance and luminous flux according to ECE standard are determined at a distance of 25 m from the light source [6]. Regardless of the processing parameters, a maximum illuminance of about 76 lx and total luminous fluxes of about 560 lm are obtained (Figure 4). An influence of the mold temperature according to the examination of the molding accuracy cannot be observed. An exception to better optical properties is seen at the testing point with a mold temperature of 190°C and an injection flow rate of 4 cm³/s. Thereby, it is probable that an LED with a higher lighting class was over-molded.
Overall, the optical properties do not completely achieve the expected qualities. This is due to impurities based on the LSR dosing system which cannot be completely cleaned of the previously processed material. 
Through the consistent use of only one material to the dosing system, these shortcomings can be significantly reduced. Experiments show that the total luminous flux can be increased and the maximum illumination by approximately 5% (Figure 4, right). Moreover, it is still possible to choose LED with higher lighting class.

Long-term Tests

As a result of previous investigations, the injection molding of LSR-optics directly onto the LED board with high molding perfection and good optical performance is possible. No influence of the high temperatures and pressure loads during the injection molding process on the optical properties was seen. Nevertheless, considering the long life cycle of a LED module it cannot be excluded that the performance of the LED is influenced later on. In order to evaluate the long-term properties, two long-term-tests were carried out regarding the standards JESD22-A104D and JESD22-A108C. On the one hand, a temperature-cycling-test between -40°C and +125°C for 1,000 cycles (half an hour per cycle) was performed in order to simulate climate influences during the use of the LED. On the other hand, an operating-life-test of the LED with the permitted maximum current of 1 A was carried out for 2,000 h in order to evaluate a damaging of the optics due to high UV- and blue light radiation as well as high temperatures. As a reference, the long-term-tests were carried out with LED boards which were not over-molded. For an economical process, the mold temperature should be as high as possible. Therefore, for the long-term tests optics with a high mold temperature (190°C) were chosen.

After the injection molding process, an annealing of the optics is recommended [7]. Therefore, the optics were annealed for 8 h at a temperature of 130°C, which lies below the recommended operating temperature of the LED board.
In order to evaluate the photometric properties the color of the emitted light of the LED was evaluated before and after injection molding, after annealing and after the testing conditions. Therefore, the CIE standard colorimetric system was used (Figure 5). Thereby, the percentage of red, blue and green of the emitted light was considered.

Figure 5. CIE standard colorimetric system [8]

The results of the long-term-tests are shown in Figure 6 and 7. Before over-molding the LED show color value x and y of around 0.33. That is why the resulting light appears white (Figure 5). After over-molding both color values are significantly reduced to 0.28 for color value x and 0.27 for color value y (Figure 5). This implicates a color shift to blue. Due to the contact of the silicone with the semiconductor chip, the difference in refraction index between the semiconductor chip and the environment becomes lower. Therefore, the total internal reflection of light with a low wavelength is reduced such that more blue light is emitted. That problem can be solved by modifying the converter (luminescent substance) of the LED chip. The converter of the used LED is designed for the transition from the semiconductor chip to air. The converter has to transform more blue light into red and green light in order to compensate for the color shift. Contrary to the over-molding, both long-term-tests as well as the annealing step before show no significant influence on the color distribution. In conclusion, a long-term use of LED with combined primary and secondary optics is possible without significant changes in the LED color.

Figure 6. Color value of the LED before and after the operating life testing
Figure 7. Color value of the LED before and after temperature cycling testing

Optimization of the Optical Function Using Diffractive Structures

Analysis of the optical properties and different long-term-test have shown that a use of the LED module with silicone optics is possible. However, regarding freedom in design, the full potential of the material LSR is not used so far. Micro- and nano-structured surfaces offer the potential to significantly improve the optical performance of the LED module. Compared to thermoplastics liquid silicone rubber has a low processing viscosity between 20-500 Pa·s (shear rate 10 s-1, 20°C) in combination with a high elasticity [9]. Therefore, the replication and demolding of micro- and nanostructures in injection molding are easily possible without needing a dynamic mold tempering.
In order to improve the optical properties of the developed LSR optics, a diffractive Fresnel structure was designed. To enable a comparison to the optics without structures, the macroscopic geometry of the optics was  kept the same and was only extended with a structured area in the center (Figure 8, left). Afterward the replication accuracy of the Fresnel structures was analyzed with a digital 3D laser scanning microscope VK-X 200, Keyence Deutschland GmbH, Neu-Isenburg, Germany. Therefore, the structure heights at the center of the optics were considered. The structured optics were molded with two different mold temperatures (130 and 160°C). The measurements show structure heights between 1.15 and 1.20 µm for both mold temperatures (Figure 8, right). The structure heights of the mold insert are in the same range. Thus, a good molding accuracy of the structures was achieved. At some points, the structure heights of the optics show slightly higher values than the mold inserts. Experiences with thermoplastic materials show, that a stretching of the structures during demolding is possible. Due to the elastic material behavior of LSR, it is more probable that higher structures of the optics are caused by errors of measurement. The measurements were carried out at room temperature. Therefore, the influences of thermal expansion could not be considered.
Figure 8. Designed optics with a diffractive Fresnel-structure (left), structure height independence of the position and mold temperature (right)

In order to evaluate the optical properties, the luminance of the LED was analyzed with LSR optics with and without diffractive structured surfaces. Therefore, the LED was placed in front of a diffusely transmitted screen and the luminance was measured with a luminance and color measuring camera LMK 98-3 Color of the company TechnoTeam Bildverarbeitungs GmbH, Ilmenau (Figure 9). Using LSR optics it can be seen, that the Lambertian light distribution of the LED is deformed elliptically (Figure 10). This shape fits the  secondary optics that is used in an automotive headlight. As a consequence, more light from the LED can be targeted on the street. The luminance distribution of the structured optics shows irregularities in the bottom right area which may be caused by flush between the optical and structured mold insert. In order to classify the optical performance, the maximum luminance was compared. Using unstructured optics the maximum luminance can be increased by 22%. Using optics with the diffractive surface it is possible to extend the maximum luminance up to 30%. However, the measured luminance and luminance distribution is only an indicator and can be used comparably between the three set-ups. Nevertheless, the analyzed luminance distribution shows, that more light can be projected on a secondary optics if the surface of the optics is modified with diffractive structures.
Figure 9. Experimental set-up for the analysis of the optical properties

Figure 10. Luminance distribution and maximum luminance Lmax of the LED without optics and unstructured and structured optics


The investigations show that the concept of an injection molded combined primary and secondary optics made of highly transparent liquid silicone rubber (LSR) has been successfully realized. In comparison of the conventional solution one optical element can be saved and therefore, time and cost intensive mounting steps can be reduced. The material and the LED can fulfill long term requirements regarding temperature and maximum current. After both long-term-tests no significant color shift was measured. However, a color shift to blue light is seen directly after overmolding. Therefore, the converter of the LED has to be modified in order to generate the desired light color output. The next step is the implementation of the concept in a serial application. Thereby, it is possible to improve the optical function using a diffractive Fresnel-structure at the surface of the LSR optics.

Credits : Christian Hopmann and Malte Röbig * Institute of Plastics Processing (IKV) at RWTH Aachen University, Aachen, Germany

No comments:

Post a Comment