Development of Laser Confocal Microscopy for Internal Defect Measurement

Transcription

1 Development of Laser Confocal Microscopy for Internal Defect Measurement Chia-Liang Yeh*, Fu-Cheng Yang, Wei-Hsiung Tsai, and Keng-Li Lin Center for Measurement Standards, Industrial Technology Research Institute, Hsinchu, Taiwan, 300, R.O.C Abstract Organic light-emitting display has flexibly deformable characteristics. However, when the foldable display is repeatedly folded and unfolded, which may cause the film between the metal electrodes and the active layers to delaminate. Greatly extend the operation lifetime is discussed with a focus on materials science and structure design. Accordingly, we have developed a nondestructive measurement that can identify potential defects such as delamination before ensuing manufacturing steps are applied. 1. Introduction The annual global lighting market is worth approximately US $100 billion. When an OLED is operated at maximum brightness (3000 cd/m - ), its lifetime is estimated to be 30,000 40,000 hours. When the brightness is reduced by one-third (i.e., to 000 cd/m - ), its lifetime almost doubles, reaching 54,000 70,000 hours. Therefore, OLEDs are excellent for applications that require low light emittance and long life [1]. At present, global lighting and panel manufacturers, as well as governments in every developed country, are investing heavily in the research and development of OLED lighting. According to an OLED lighting market report released by NanoMarkets, the revenues for OLED lighting could exceed US $1.3 billion in 018 (Fig. 1) []. Figure 1. OLED growth forecast [] The key challenge with foldable displays is that they must be able to endure repeated folding and unfolding and avoid delamination. The strength of this feature in OLEDs is dependent on the material and structural design used. When delamination occurs, the starting position of the error must be traced in order to improve the design. Here, we develop an optical method of measurement that primarily uses laser confocal technology to address the limitations caused by poor backlight, a poor signal-to-noise ratio, and an insufficient dynamic range. This method allows us to measure the multilayer film structure of an OLED, and thus solve the lack of device problem for measuring film delamination.. Research method In the proposed optical system, a circular pinhole is placed at the position where the focal points of the objective lens conjugate, so that only the light at the focal point can penetrate; the light is reflected on the sample surface and returned along the same optical path, diverged by the beam splitter, and then converges on the pinhole [3]. Notably, the optical system can only obtain information from the focal point because most of the reflected light from the off-focus position was blocked at the pinhole. This means that a confocal (i.e., one focal point) optical system can measure the three-dimensional shape of an object, because the depth of focus is very narrow; traditional optical systems receive unfocused signals, which causes the depth of field to be relatively larger. Additionally, compared with traditional optical microscopes (in which out-offocus signals are superimposed), confocal optical systems block reflected light from areas outside the pinhole, forming a clear image with excellent contrast. Although the diameter of the pinhole should be smaller than the spot light generated by spot diffraction in confocal optical systems, pinholes that are smaller could significantly affect luminous flux during system optimization. Because resolution enhancement is very limited, the ideal method is to validate the pinhole diameter [4] through the sensitivity of the light source receiver. Confocal microscopes can filter the out-of-focus light above and below the focal point of the sample through the pinhole. When objects are viewed in the microscope, the signal is generated from the entire thickness of the test object; thus, most of it cannot focus. However, confocal microscopes eliminate the out-of-focus signals through the confocal pinhole positioned in front of the image plane. The pinhole then acts as a space filter, and only permits focused areas to be viewed, eliminating the light above and below the focal point of the sample in the final image. The confocal principle is illustrated in Fig. [5].

2 ε = 0.55 λ NA (1) Figure. Confocal principle [5] In this study, a dual-channel confocal optical system was developed, and lead zirconate titanate doped with zirconium impurities (PbZrxTi1-xO3 or PZT) was used on the objective lens as a z-axis scan. When the objective lens was at a specific height, Pinhole 1 was placed at the confocal position of the object, and Pinhole was placed using PZT in front of and behind the confocal point to reciprocate the sine function. Avalanche photodiodes (APD) were used as detectors. Both APD 1 and APD detected the confocal light signals, and the transimpedance amplifier (TIA) was inputted using an AC coupling mode. Then, a mixer was inputted and the PZT of Pinhole was initialized for the synchronous modulation of the signal mixer, which converted the original confocal signal into a differential signal. This successfully removed the limitations of background and noise. Figure 3 shows an outline of the structure of our proposed system. Figure 3. Dual-channel confocal optical path According to the Rayleigh criterion [6], resolution is dependent on the Airy disk. Due to the pinhole effect, the lateral resolution power of a confocal microscope is superior to that of a traditional microscope; as shown in Fig. 4, the lateral resolution ε is: where λ is the wavelength, NA is the numerical aperture of the lens, and the three-dimensional resolution of confocal microscopy mainly comes from the beam at the axial depth of the field. The ε axial resolution FZC can be expressed as follows [7] : λ ε FZC = 1.8 NA n 1 1 n () Figure 4. Confocal optical system The present study used a 405 nm laser, NA 0.9 lens, 44.4 nm lateral resolution, and nm z- axis resolution. Based on the theory of thin-film interference, the intensity I(λ) can be expressed as follows when using different wavelengths to measure single-layer thin films [8] : [ δ ( )] I ( λ) = I1( λ) + I ( λ) + I1( λ) I ( λ) cos λ (3) where, λ is the wavelength used for the measurement, I 1(λ) is the intensity of the reflected light on the upper surface of the film, and I (λ) is the intensity of the reflected light on the bottom surface of the film. Additionally, δ(λ) is the phase difference between the two intensities, expressed as 4πn ( λ)t δ ( λ) = λ (4) where n(λ) is the refractive index of the material and t is the thickness of the film. Figure 5 shows the reflectance distribution of each layer of material in the OLED display under different wavelengths. The materials with high reflectivity were primarily detected in the metal layer, and because of the large difference in the refractive index, there was greater reflectivity between the interfaces. Notably, spot size and measurement resolution have a positive correlation when wavelength is considered. A 405 nm laser was adopted for the measurement.

3 peak, and the curve overlapping and curve fitting for each peak was resolved. The maximum composite waveform conformed to the actual waveform, which thus revealed the error correction function. Figure 6. (a) Graph of the reflectivity of the signal, which corresponds to the depth of each layer of the OLED display. (b) Graph of the reflectivity of the signal, which corresponds to the depth of the main light-emitting layer of the OLED display. (c) Graph of the reflectivity of the signal layers after coupling, as measured by the confocal system. Figure 5. (a) Reflectance distribution of each layer of material of the OLED display under different wavelengths. (b) Reflectivity of each layer of material of the OLED display under 405 nm wavelengths. Upon examining the reflectance and thickness distribution of each layer, it was clear that the reflected signal mainly appeared on the polyimide (PI) of the upper layer and on the electrode layer at the bottom. Using a signal with a wave packet of relative intensity and the same full width at a half maximum Gaussian distribution, the signal distribution of the confocal system was simulated. The stacked film layer was processed using the signal superposition method, and as depicted in Figure 6, the measured signal uses signal processing and separation. The Levenberg Marquardt algorithm,[9] the training speed of which can be increased using the dynamic learning rate, was used for the fitting calculation process. Specifically, the learning rate of each gradient direction can be forecasted with the Hessian matrix, which can greatly enhance the speed and accuracy of the calculations when combined with the advantages of the Gauss Newton algorithm and gradient descent method. Here, different Gaussian and Lorenz distribution functions were used to eliminate signal error to fit the different curves. Finally, an iterative method was adopted to determine the least squares solution within the range of the corresponding parameters for each Because the difference in the reflectivity of each layer of the OLED display was not substantial, and because most of their thickness was very thin, the measurement of the inner layer shape required multiple interface signals to be measured. Although the use of laser confocal technology can eliminate the difficulties of analyzing dual wave packet and multilayer film interference, such technology is still affected by background and noise when a confocal signal is used to measure the inner layer shape. This is due to the limitations of confocal signals that are caused by diffraction limits, which face mutual interference between multiple interfaces and confocal signals in the axial direction, and because the inner layer signal is relatively weak. In the present study, we used a heterodyne and a differential method to solve the characteristic limitations in the proposed confocal system. Specifically, we mixed two confocal signals and eliminated noise through a low-pass filter to reduce interference. 3. Research results Figure 7 shows the system architecture, which was designed based on the optical path diagram. The collimated laser was directed to the test object through the objective lens and the reflected signal, after being divided, was directed into the two APDs via the collimator lens. Notably, there is a pinhole in front of the APD. Moreover, the z-axis position that

4 was measured by adjusting the objective lens and the infinity corrected optical system formed the visual system. and the difference in reflectivity was minimal, the differential method could separate the subtle changes in the reflected wave packet to identify various positions. Figure 7. Confocal system architecture and component location diagram As illustrated in Figure 8, a spatial filter was added to the end of the laser to filter spatial noise, which enhanced the quality of the laser beam and the beam quality at the end of the objective lens. Additionally, an optical isolator was added to stabilize the signals. By limiting the direction of the beam, the isolator prevented excessive reflected light from entering the laser cavity and affecting signal stability. Figure 9. Subtle changes in the reflected wave packet that reveal various positions of interference problems between multiple interfaces and confocal signals Finally, Figure 10 presents the actual measurement of the multistructured transparent display. Notably, both delaminated and nondelaminated samples were selected for the test. Using the laser confocal system, different signals were detected at the fourth layer of the delaminated and nondelaminated samples. The delamination distance was 4 μm. Figure 8. Comparison of the beam quality of a 405 nm laser before and after passing through the spatial filter A comparison of the results of the laser beam passing through the five layers is shown in Figure 9. When two layers of film were positioned closely,

5 Pearson Education Chapter 6, Problems (007) [7] Zenhausern F, Martin Y, Scanning interferometric apertureless microscopy. Science 69: (1995) [8] Stavenga, D. G. "Thin Film and Multilayer Optics Cause Structural Colors of Many Insects and Birds". Materials Today: Proceedings. 1: 109(014) [9] K. Levenberg, A method for the solution of certain problems in least squares, Quart. Appl. Mach., vol., pp , (1944) Figure 10 Results of delaminated and nondelaminated samples measurement 4. Conclusions The structure of a foldable OLED display must be able to endure repeated folding and unfolding and avoid delamination, features which are dependent on the materials used and structural design. When a test fails, the source of the delamination must be traced in order to correct the design. In this study, we developed an optical measurement method that is based primarily on laser confocal technology to eliminate the problems with analyzing dual wave packets and of multilayer film interference. Paired with the heterodyne and differential methods, the proposed method effectively solved the problems of confocal signals that were limited (in terms of axial direction) by mutual interference between multiple interfaces and confocal signals. In short, the proposed method facilitates measurement of the structure of multilayer films, and resolves the problem of the lack of a device for tracing film delamination. References 1. S. T. Wu and D. K. Yang, Reflective Liquid Crystal Displays, (Wiley-SID, 001). [1] Michael Royer, Lumen Maintenance and Light Loss Factors : Consequences of Current Design Practices for LEDs DOI : / (014) [] Nano Markets. Accessed Jul from: (015) [3] Marvin Minsky, Filed in 1957 and granted US (1961) [4] M. Minsky, Memoir on Inventing the Confocal Scanning Microscope, Scanning,10 (1988), pp [5] Fellers TJ, Davidson MW. "Introduction to Confocal Microscopy". (007). [6] Carroll, Bradley W. & Ostlie, Dale A, An Introduction to Modern Astrophysics, nd Edition,