Open Access
Add to BibSonomy
Abstract
This paper presents the mid-field model for an ultraviolet C light emitting diode (UVC LED) of wavelength around 275±5 nm by comparison of the 2-dimension (2-D) gray-level image captured from a mono-CMOS sensor and simulated irradiance pattern. Because of UVC light, we propose using a fluorescent film to absorb UVC light and re-emit visible light so that the 2-D image could be captured. The analysis and calibration to obtain accurate gray level of image are performed. Finally, we achieve the mid-field model with high accuracy. Furthermore, this model is also applied for dome lens design and then compares the performance with fabricated samples in measurement to expertise its validity.
© 2021 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
For many years ago, we have known about the crucial roles of ultraviolet (UV) light in sterilization [1, 2]. UVC irradiation with a range of wavelengths from 100-280 nm has been considered to be the effective condition for inactivating the DNA and RNA of microorganisms [3]. In comparison with traditional lamps, UVC light-emitting diode (LED) has more advantages such as lower energy cost, long-life time, low power consumption, environmental benefit and compact size [4,5]. Applications of UVC LED to disinfect microorganisms on surface, in air and in water have been reported in numerous studies [6–8]. Recently, several novel germicidal applications have been recommended including using UVC light to disinfect Coronavirus such as creating disinfection system for whole-room and UVC LED boxes to disinfect the personal items, etc. [9–12]. Understanding the potential of UVC LED, the study in UVC LED at a wavelength of around 275 nm is performed in this paper. If we achieve a very accurate light source model, it can be used for lighting design to fabricate the design products as required. Therefore, a light source model for a UVC LED is considered a necessary step before proceeding to design a particular application [13,14].
Generally, there are three working zones of a LED: near-field, mid-field, and far-field [15]. In the mid-field region, the light field varies from one distance to another because of the chip dimension or the encapsulated structure. Therefore, the optical model for extended light source is verified in the mid-field region, where the farthest distance is 10 times the largest lateral dimension of chip of package [15,16]. To achieve the optical model for an LED, we create the model with the same parameters with LED sample and compare the simulated light distributions with those corresponding measurements at different mid-field distances. Basically, there are two methods to approach the light source model, the first method is comparison of the one-dimension (1-D) light intensity and the second method is comparison of the two-dimension (2-D) irradiance pattern. In the first method which obtains the light source model by measuring the 1-D light intensity, we use a power meter connected with a detector to receive flux over steradian by set up the LED rotates around itself from $-$ 90 to 90 degrees or reverse, the detector rotates around LED. Numerous reports approached the precise optical model by following this method [17–19]. Nevertheless, one of the challenges of this method is that the space between LED and detector must be large enough so that the LED or detector can be easily rotated. Otherwise, we could not apply this method if the emitting area is not large enough so that there will be difficulty in rotation at their mid-field distance, as described in Fig. 2(a). In the 2-D irradiance pattern measurement of the second method, the system consists of a quasi-Lambertian scattering surface to scatter the light emitted by LED into the CCD or mono-CMOS sensor [15]. However, UVC light is invisible light, to capture the light pattern emitted from light source, a fluorescent film was proposed to absorb the incident UVC light and re-emit of visible light. One of the important optical properties of fluorescent materials is that when irradiated with short-wavelength light, it re-emits light at a longer wavelength isotropically [20–22]. In this experiment, we used a fluorescent film which structure consists of two layers, one is a fluorescent pigment layer coated with epoxy resin mixed phosphor, where receives incident UVC light and the other is PET layer, as shown in Fig. 1 [23], a total thickness of about 40 μm. For different types and energy of UVC LEDs, an appropriate thickness of the fluorescent film to be selected is also different. In the case of the fluorescent film is so thin that we cannot clearly observe the visible light through it, we can overlap several pieces so that the light pattern of UVC after passing through it can be clearly observed in visible light. Furthermore, used film pieces are regularly replaced with new ones to maintain good quality. The fluorescent film could be easily moved close to light source during experiment as Fig. 2(b), so we are able to analyze the 2-D irradiance in mid-field distance of the light source. The calibration is made to ensure that the gray-level by CCD or CMOS virtually corresponds to irradiance on the fluorescent plate. To evaluate how many percent of the similarity between simulation and experiment, we apply the normalized cross-correlation (NCC) value. If the NCC reaches 99.5% or above, we are able to achieve the precise optical model [15,24]. Consequently, for UVC LEDs with a small lateral size which the mid-field distance is not large enough to measure the 1-D light intensity, we propose using fluorescent film which absorbs UV-C light and re-emits visible light so that the 2-D irradiance light pattern could be captured by a CCD or CMOS sensor, and then we compare it with the corresponding simulation in mid-field distance.
Fig. 1. Fluorescent film used for the experiment; (a) real 2-D image of film and (b) description of structure of film [23].
Fig. 2. (a) Description of experimental setup using detector to measure the light intensity distribution as angle, (b) detector replaced by fluorescent film in novel method.
In this paper, we present a light source model for a UVC LED of wavelength around 275 nm by comparison of the 2-D irradiance pattern. A suitable fluorescent film is collected and used to obtain a 2-D light pattern. The analysis and calibration in several experiments are also performed to ensure that gray level of image captured by camera corresponds to the irradiance on the fluorescent film.
2. Light source model for a particular UVC LED
2.1 Methodology
In this work, we perform the optical model for specific UVC LED which the light source operated at a wavelength of 275${\pm}$ 5 nm and located in the cavity with cover by window quartz as shown in Fig. 3(a). The size of emitting area is 1.2 mm ${\times}$ 1.2 mm ${\times}$ 0.4 mm as described in Fig. 3(b). The farthest mid-field of this light source is 10 times the largest dimension of chip, which means about 12 mm in this case, this distance is too difficult for rotation in the 1-D intensity measurement method. Hence, using fluorescent film to obtain the 2-D irradiance is applied. We make the optical model with the same parameters as a real sample including the size of all elements, weighting factor of emitting area, absorption coefficient and reflectivity as shown in Fig. 3(c) without quartz window and Fig. 3(d) with a quartz window. Then, the 2-D irradiance pattern is simulated through a Monte-Carlo ray-tracing by ASAP program [25] and compared with experiments in different selected mid-field distances.
Fig. 3. (a) UVC LED sample used for experiment, (b) description for dimension of UVC LED die, optical model for UVC LED without (c) and with (d) quartz window.
The measurement system consists of the fluorescent film to absorb UVC light emitted from LED and the mono-CMOS sensor used to capture the 2-D light pattern at different distances in mid-field. In this case, owing to the technical parameters of the mono-CMOS sensor, we located it 100 mm far from fluorescent film to captured the quality image. We defined $D$ is the distance between the top surface of LED and fluorescent film in experiment, and $D$ distance is measured from the LED tip in simulation. The UVC LED was moved to different distances far away from fluorescent film at 5 mm, 8 mm and 10 mm, respectively, these distances are considered in mid-field region. The simulation setup was also made corresponding to experiment, the description as shown in Fig. 4. However, the incident UVC light comes in fluorescent film as irradiance, the UVC light then is re-emitted in visible light and becomes light radiance. The radiance of light pattern then is captured by mono-CMOS sensor. The comparison will be made if the irradiance incidents the fluorescent film and radiance scattered by the film are virtually proportional to each other. Therefore, the appraisal of the proportionality of irradiance on the incident surface and radiance on the other surface of fluorescent film is necessary. Furthermore, the location of CMOS sensor causes in-correction for gray level at positions where tilt with normal of CMOS sensor as angle due to cosine third or fourth law as shown in Fig. 4(a). Thus, calibration for gray level of image captured by CMOS is necessary before comparing its performance with corresponding simulation. The issues are respectively presented in the next two sections.
Fig. 4. (a) Description for the experiment setup to capture light pattern by CMOS sensor, and (b) simulation setup to obtain the irradiance distribution. $D$ is the distance between top surface of LED and fluorescent film in experiment, and the distance $D$ is measured from the LED tip in simulation.
2.2 Appraising the proportionality of radiance and irradiance on the surface of fluorescent film
As mention above, appraising the proportionality of irradiance on the incident surface of the film and radiance on the re-emitting surface is required. We make the setup so that the re-emitted light forms the small light spot by using two irises. The incident light is fixed while we are moving up and moving down the film, the experimental setup as shown in Fig. 5(a) and the photo of light spot as shown in Fig. 5(b). During moving, we obtain the same amount of incident light in different positions of the film. If the light spots of all selected positions to test on fluorescent film are virtually similar to each other, the proportionality of irradiance and radiance on its surface is demonstrated. To simplify, we compare the central gray level of light spots of several positions on film. Seven selected positions are marked and described as shown in Fig. 6(a), one of seven light spots by mono-CMOS is displayed as Fig. 6(b), and the comparison result indicates that the gray level at the center of seven light spots are almost similar to each other as illustrated in Fig. 6(c). Therefore, the irradiance on the incident surface of fluorescent film is proportional to the radiance on its re-emitting surface.
Fig. 5. (a) Experimental setup to appraise the proportionality of radiance and irradiance on surface of the film and (b) the photo of light spot on the fluorescent film.
Fig. 6. (a) Marked and tested seven positions, (b) the specific light spot image by mono-CMOS for one of seven positions, and (c) comparison of the gray level in the center of the seven light spots.
2.3 Calibration equation for gray level
The gray level of central position where lies in the normal with CMOS sensor is considered accurate value while others are inaccurate. To figure out the calibration equation for gray level of all positions as angle, the experimental setup is established as shown in Fig. 7. The idea is to create a light spot on the fluorescent film by using two irises to narrow incident light emitted from UVC LED similar to the light spot as shown in Fig. 5(b) and move the CMOS sensor to different positions as the direction of perpendicular to its own normal, the gray level on the center of light spot is collected at many positions correspond to different angles. Angle $\theta$ is defined as the angle between the central point of CMOS sensor and the normal direction of light spot. With the same exposure time of CMOS sensor, so we are able to collect the gray level of the central point on light spot at different angle $\theta$. The results indicate that the central gray level attenuates when the magnitude of $\theta$ is increased. Gray level of central point at several positions of CMOS sensor are compared and fitted as shown in Fig. 8, the fitted equation also is figured out as
where E, ${E_0}$ is the gray level of central position on light spot by CMOS sensor located angle $\theta $, $\theta = {0^ \circ }$, respectively. However, this experiment is established to figure out the calibration equation, after that, to apply it for calibration of an image, the new definition is significant as where $B$ is gray level captured by CMOS sensor, and ${B_0}$ is the real gray level value, without angular error. The images captured by CMOS sensor are calibrated by using Eq. (2). Namely, we have the gray level matrix of image, each gray level position corresponds to angular position, to approach the gray level as accurately as possible, ${B_0}$, while the image matrix by CMOS sensor is $B$, we divide $B$ for the right side of Eq. (2).Fig. 7. Experimental setup to capture the spot light at different angle by moving CMOS sensor at different angle $\theta $.
2.4 Results
The experimental setup using CMOS to capture the light pattern on fluorescent film emitted by UVC LED as shown in Fig. 9(a), the photos for light pattern on fluorescent film at $D = $ 5 mm, $D = $ 8 mm and $D = $ 10 mm as shown in Fig. 9(b), (c) and (d), corresponding. The image after captured by mono-CMOS is calibrated using Eq. (2), as shown in Fig. 10(a) then compared with the 2-D irradiance pattern in simulation at $D = $ 5 mm by normalization as shown in Fig. 10(b), with dimension as 40 mm ${\times} $ 40 mm. The horizontal and vertical distribution are also compared as shown in Fig. 10(c) and (d), corresponding. The NCCs are calculated as 99.8% for both horizontal and vertical. Similarly, distance $D = $ 8 mm and $D = $ 10 mm are analyzed and compare as shown in Fig. 11 and 12, respectively. For three selected distances in mid-field, the results indicate that the light source model is precise enough to apply for lighting designs.
3. Application of light source model for dome lens design
To evaluate the validity of the light source model, we apply the model for dome lens design which structure is as shown in Fig. 13(a), (b), dome lens design on the cavity of LED in simulation as shown in Fig. 13(c), this design was manufactured such Fig. 13(d), and lens sample attaches on cavity of LED for measurement as shown in Fig. 13(e). The experiment is arranged to capture the 2-D gray-level image as shown in Fig. 14(a), and the photos for light pattern on fluorescent film at $D = $ 8 mm, $D = $ 10 mm and $D = $ 15 mm as shown in Fig. 9(b), (c) and (d), corresponding. We compare the normalization of 2-D gray-level image after calibration using Eq. (2) and 2-D simulated irradiance pattern at distance $D = $ 8 mm (mid-field), $D = $ 10 mm (mid-field), $D = $ 15 mm (far-field) with dimension as 40 mm ${\times} $ 40 mm. The results are shown in Fig. 15, where NCC results at distances $D$ = 8 mm (mid-field) are 99.6% as horizontal and 99.7% as vertical, at $D = $ 10 mm (mid-filed) are 99.7% as horizontal and 99.8% as vertical, $D = $ 15 mm (far-field) are 99.9% as horizontal and 99.8% as vertical. Fortunately, the comparison of both simulation and experiment are virtually matched to each other. Therefore, we again indicate the precision and validity of light source model we achieved in this work.
Fig. 9. (a) Experimental setup using CMOS to capture the light pattern on fluorescent film emitted by UVC LED. The photos for light pattern on fluorescent film at (b) $D = $ 5 mm, (c) $D = $ 8 mm and (d) $D = $ 10 mm.
Fig. 10. Comparison between experiment and simulation at distance $D = $ 5 mm; (a) light pattern after calibration and normalization in experiment, (b) light pattern after normalization in simulation; Comparison of the irradiance distribution in simulation and gray level distribution in experiment as (c) horizontal and (d) vertical.
Fig. 11. Comparison between experiment and simulation at distance $D = $ 8 mm; (a) light pattern after calibration and normalization in experiment, (b) light pattern after normalization in simulation; Comparison of the irradiance distribution in simulation and gray level distribution in experiment as (c) horizontal and (d) vertical.
Fig. 12. Comparison between experiment and simulation at distance $D = $ 10 mm, (a) light pattern after calibration and normalization in experiment, (b) light pattern after normalization in simulation; Comparison of the irradiance distribution in simulation and gray level distribution in experiment as (c) horizontal and (d) vertical.
Fig. 13. (a) Two-dimensional contour, (b) three-dimensional structure of the dome lens design, (c) dome lens design attached on cavity, (d) dome lens sample, and (e) dome lens attached on LED sample.
Fig. 14. (a) Experimental setup using CMOS to capture the light pattern on fluorescent film emitted by UVC LED attached to dome lens. The photos for light pattern on fluorescent film at (b) $D = $ 8 mm, (c) $D = $ 10 mm and (d) $D = $ 15 mm.
Fig. 15. NCC results and the light patterns in both simulation and measurement at $D = $ 8 mm, $D = $ 10 mm, $D = $ 15 mm, where S and M are abbreviations of simulation and measurement, corresponding.
4. Conclusion
The effectiveness in deactivation of the microorganisms and the compact size are two remarkable advantages of UVC LED when applying to fabricate the sterilization products. Before start of manufacturing for a specific product, the design is the important procedure to predict the optical behavior. To achieve the samples virtually similar to lighting design, owning an accurate optical model is required. Hence, the light source model in mid-field region is really necessary before starting a lighting design.
There are two methods to approach the mid-field model, one is the comparison of the 1-D light intensity and the other is comparison of the 2-D light pattern in both experiment and simulation. However, to measure the 1-D light intensity, we use a power meter connected with detector to receive flux over steradian by rotation, this leads to requirement about the spacing enough between LED and detector. Therefore, it is challenge when measuring the 1-D intensity for LEDs which mid-field distance is limit. In this paper, we apply the method by comparison of the 2-D light pattern for a UVC LED with the farthest mid-field distance about 12 mm that is difficult to proceed the measuring 1-D light intensity by rotation. Because UVC light is invisible, we propose using the fluorescent film to absorb UVC light and re-emit visible light, and using a mono-CMOS sensor to capture light pattern. However, the incident UVC light comes in fluorescent film as irradiance, the UVC light then is re-emitted in visible light as radiance. The comparison is effective if the irradiance incidents the fluorescent film and radiance scattered by the film are virtually proportional to each other. Therefore, the appraisal of the proportionality of irradiance on the incident surface and radiance on the other surface of fluorescent film was proceeded. Furthermore, the location of CMOS sensor causes an inaccuracy for the gray level at positions where tilted normally by CMOS sensor as angle due to cosine third or fourth law. Thus, it is necessary to figure out the calibration equation for the gray level of the image captured by the CMOS.
In conclusion, the proportionality of irradiance on the incident surface and radiance on the other surface of fluorescent film was demonstrated. Calibration equation for gray level of image as angle was also figured out. Then, the light source model was approached by comparing the 2-D gray-level pattern after calibrating using Eq. (2) and the 2-D irradiance pattern in simulation as normalization. The NCC results of several selected distances in mid-field are 99.8% at $D = $ 5 mm, and 99.9% at both $D = $ 8 mm and $D = $ 10 mm, which means the light source model that we achieved is precise. Furthermore, this model was applied for the dome lens design to appraise how it valid is. The dome lens design was compared with lens sample about optical behavior when it was attached to LED. We also obtained the 2-D gray-level pattern by using fluorescent film and mono-CMOS, then calibrating by applying Eq. (2) and comparing with the 2-D irradiance pattern in simulation as normalization. Indeed, the performance of using the achieved model to design dome lens practically matched with measurement. The NCC results at distances $D = $ 8 mm (mid-field) are 99.6% as horizontal and 99.7% as vertical, at $D = $ 10 mm (mid-filed) are 99.7% as horizontal and 99.8% as vertical, $D = $ 15 mm (far-field) are 99.9% as horizontal and 99.8% as vertical. Consequently, we propose the method of using fluorescent film to achieve the precise mid-field model for a UVC LED.
Funding
Ministry of Science and Technology, Taiwan (109-2221-E-008-087-MY2, 109-2622-E-008-014-CC2, 109-2622-E-008-026).
Acknowledgments
The author would like to thank Breault Research Organization (BRO), Inc. for sponsoring ASAP software program.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. K. C. Grant, “Sterilization of polluted water by ultra-violet rays,” Eng. News 64, 275 (1910).
2. W. F. Wells, M. W. Wells, and T. S. Wilder, “Viability of B. coli exposed to ultra-violet radiation in air,” Science 82(2125), 280–281 (1935). [CrossRef]
3. T. Bintsis, E. Litopoulou-Tzanetaki, and R. K. Robinson, “Existing and potential applications of ultraviolet light in the food industry – a critical review,” J. Sci. Food Agric. 80(6), 637–645 (2000). [CrossRef]
4. A. Zukauskas, M. S. Shur, and R. Caska, “Introduction to Solid-state Lighting,” John Wiley & Sons, New York (2002).
5. D. A. Steigerwald, J. C. Bhat, D. Collins, R. M. Fletcher, M. O. Holcomb, M. J. Ludowise, P. S. Martin, and S. L. Rudaz, “Illumination with solid state lighting technology,” IEEE J. Select. Topics Quantum Electron. 8(2), 310–320 (2002). [CrossRef]
6. S. Rattanakul and K. Oguma, “Inactivation kinetics and efficiencies of UV-LEDs against Pseudomonas aeruginosa, Legionella pneumophila, and surrogate microorganisms,” Water Res. 130, 31–37 (2018). [CrossRef]
7. R. Nishisaka-Nonaka, K. Mawatari, T. Yamamoto, M. Kojima, T. Shimohata, T. Uebanso, M. Nakahashi, T. Emoto, M. Akutagawa, Y. Kinouchi, and T. Wada, “Irradiation by ultraviolet light-emitting diodes inactivates influenza a viruses by inhibiting replication and transcription of viral RNA in host cells,” Journal of Photochemistry and Photobiology B: Biology 189, 193–200 (2018). [CrossRef]
8. D. K. Kim and D. H. Kang, “UVC LED irradiation effectively inactivates aerosolized viruses, bacteria, and fungi in a chamber-type air disinfection system,” Appl Environ Microbiol 84(17), e00944 (2018). [CrossRef]
9. H. Inagaki, A. Saito, H. Sugiyama, T. Okabayashi, and S. Fujimoto, “Rapid inactivation of SARS-CoV-2 with deep-UV LED irradiation,” Emerging Microbes & aInfections 9(1), 1744–1747 (2020). [CrossRef]
10. K. Bedell, A. H. Buchaklian, and S. Perlman, “Efficacy of an automated multiple emitter whole-room ultraviolet-C disinfection system against coronaviruses MHV and MERS-CoV,” Infect. Control Hosp. Epidemiol. 37(5), 598–599 (2016). [CrossRef]
11. D. M. Ozog, J. Z. Sexton, S. Narla, C. D. Pretto-Kernahan, C. Mirabelli, H. W. Lim, I. H. Hamzavi, R. J. Tibbetts, and Q. S. Mi, “The effect of ultraviolet C radiation against different N95 respirators inoculated with SARS-CoV-2,” Int. J. Infect. Dis. 100(9), 224–229 (2020). [CrossRef]
12. R. J. Fischer, D. H. Morris, N. van Doremalen, S. Sarchette, M. J. Matson, T. Bushmaker, C. K. Yinda, S. N. Seifert, A. Gamble, B. N. Williamson, and S. D. Judson, “Effectiveness of N95 respirator decontamination and reuse against SARS-CoV-2 virus,” Emerg. Infect. Dis. 26(9), 2253–2255 (2020). [CrossRef]
13. M. S. Kaminski, K. J. Garcia, M. A. Stevenson, M. Frate, and R. J. Koshel, “Advanced topics in source modeling. In Modeling and Characterization of Light Sources,” Int. S. Photo-Opt. 4775, 46–57 (2002). [CrossRef]
14. H. Zerhau-Dreihoefer, U. Haack, T. Weber, and D. Wendt, “Light source modeling for automotive lighting devices. In Modeling and Characterization of Light Sources,” Int. S. Photo-Opt. 4775, 58–66 (2002). [CrossRef]
15. C. C. Sun, T. X. Lee, S. H. Ma, Y. L. Lee, and S. M. Huang, “Precise optical modeling for LED lighting verified by cross correlation in the midfield region,” Opt. Lett. 31(14), 2193–2195 (2006). [CrossRef]
16. C. C. Sun, W. T. Chien, I. Moreno, C. C. Hsieh, and Y. C. Lo, “Analysis of the far-field region of LEDs,” Opt. Express 17(16), 13918–13927 (2009). [CrossRef]
17. C.C. Sun, Y.C. Lo, C.C. Tsai, X.H. Lee, and W.T. Chien, “Anti-glare LED projection lamp based on an optical design with a confocal double-reflector,” Opt. Commun. 285(21-22), 4207–4210 (2012). [CrossRef]
18. H. J. Lin, C. C. Sun, C. S. Wu, X. H. Lee, T. H. Yang, S. K. Lin, Y. J. Lin, and Y. W. Yu, “Design of a Bicycle Head Lamp Using an Atypical White Light-Emitting Diode with Separate Dies,” Crystals 9(12), 659 (2019). [CrossRef]
19. C. S. Wu, K. Y. Chen, X. H. Lee, S. K. Lin, C. C. Sun, J. Y. Cai, T. H. Yang, and Y. W. Yu, “Design of an LED Spot Light System with a Projection Distance of 10 km,” Crystals 9(10), 524 (2019). [CrossRef]
20. K. Barnard, “Color constancy with fluorescent surfaces,” in Color and Imaging Conference, (Society for Imaging Science and Technology, 1999), pp. 257–261.
21. B. Valeur and M. N. Berberan-Santos, Molecular fluorescence: principles and applications, (John Wiley & Sons, 2012).
22. T. Treibitz, Z. Murez, B. G. Mitchell, and D. Kriegman, “Shape from fluorescence,” in European Conference on Computer Vision, (Springer, 2012), pp. 292–306.
23. Ref. https://hsdtc.com/fluorescent-film/
24. J. P. Lewis, “Fast template matching,” Proc. Canad. Imag. Proc. 19, 120–123 (1995).
25. ASAP, / http://www.bro.com/S.
