Open Access
Add to BibSonomy
Abstract
An infrared (IR) scene generation chip based on multi-layer thin film was demonstrated. Infrared radiation was efficiently converted from visible light energy through chip absorption. We proposed that the contradictive material parameter requirements aroused by high-spatial resolution and high-frame rate of the dynamic IR scene generator could be successfully solved by a periodical microstructure fabrication. The theoretical simulation results based on a semi-one-dimensional heat transfer model illustrated that the spatial resolution of the generated IR scene was improved using a microstructure with a low contact area ratio and high fill factor, while the thermal decay time of the chip was kept unaffected. Two infrared scene generation chips with different microstructures were fabricated by lithography and site-selective deposition. Measured by non-contact thermography, the chip with low microstructure contact area ratio (0.17) showed a higher spatial resolution (13.2 lp/mm) than that with high microstructure contact area ratio of 0.46 (8.0 lp/mm). Moreover, both chips had the same thermal decay time of 20 ms. The experimental results indicated that the proposed method is an effective and economic way to improve the performance of infrared scene generation chips.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
To test the performance of modern infrared detection systems, infrared characteristics of interested targets need to be simulated during the development of the systems [1,2]. Infrared scene generator is a simulation device, which receives infrared target data from computer and generates the corresponding infrared scene. It is widely used in both military and civilian fields [3–5]. Infrared scene can be generated either by direct infrared radiation emission or modulation of a homogenous infrared source. Typical direct radiation devices include resistor array, laser diode array, and infrared thin film transducer [6–9]. Furthermore, frequently used radiation modulation devices are digital micro-mirror array (DMD), liquid crystal spatial modulator, and MEMS optical attenuator array [10–12]. Although these devices have been well developed, there are still certain application limitations. Resistor array has limited array size because of its complicated fabrication and packaging process. MEMS optical attenuator array has the same problem as a resistor array; besides a low frame rate. Compared to a resistor array and the MEMS optical attenuator array, a large array size could be realized by DMD, liquid crystal spatial modulator, and laser diode array. However, a DMD usually suffers from low spatial resolution and low contrast in far-infrared band induced by illumination light diffraction; liquid crystal spatial modulator has unsatisfactory performance in mid- and far-infrared range due to the material limitation, and laser diode array requires complicated optical systems to achieve satisfactory spatial-intensity uniformity and image quality [13–15].
Compared to the aforementioned technologies, infrared thin film transducer is considered as a promising technology because both large array size and wide radiation spectral band can be obtained. This is due to the recent development of thin film fabrication technologies. Infrared thin film chip is essentially an optical-thermal-optical transducer based on heat transfer and thermal radiation. Its properties depend strongly on the thermophysical parameters of the film. High-spatial resolution requires low thermal conductivity; however, high-frame rate needs high-thermal conductivity. Therefore, it is very difficult to obtain high-frame rate and high-spatial resolution, simultaneously.
To solve this problem, we fabricated in-plane microstructures on the film to mitigate the thermal diffusion, which caused the deterioration of the spatial resolution of the generated IR image. The experimental results showed that the spatial resolution increased by more than two times compared to that in a non-structured film without affecting the thermal decay time. A theoretical model was proposed to optimize the parameters of the microstructure. The experimental results coincided well with the theoretical simulations. It reveals that materials with high thermal conductivity can be used to obtain high-frame rate of the IR dynamic scene, and by fabricating the microstructure on the film, one can obtain satisfactory spatial resolution as well.
2. Working principle and microstructure design
The infrared thin film transducing chip was composed of multiple thin layers, as shown in Fig. 1. It consisted a substrate layer, an adhesive layer, and an absorption/radiation layer. The substrate layer was Polyimide (PI). It supported the upper layers and was stable in a wide temperature range (from 20 $^\circ \textrm{C}$ to 300 $^\circ \textrm{C}$). The adhesive layer (chromium) was used to enhance the bond between the substrate layer and the absorption/radiation layer. The absorption/radiation layer was metal black, which had high absorption coefficient at incident light wavelength and high emissivity for infrared wavelengths, to ensure an efficient conversion from visible to infrared radiation.
When the thin film chip was illuminated by the incident light, it absorbed the incident light energy and heated up. If the incident light intensity had spatial distribution, for example: a grayscale image, a corresponding temperature distribution on the chip was built. The chip then generated an infrared radiation field with the same intensity distribution (a thermal image) as the incidence. To obtain a high-frame rate, a thin film with high-thermal conductivity was required. However, the spatial resolution would be deteriorated by strong in-plane thermal diffusion. To solve this problem, we proposed to use in-plane microstructures to reduce the thermal diffusion.
Figure 2 shows the microstructure; a part of the film was removed by lithography and etching. Two parameters, contact area ratio (ηc) and fill factor (ηf), were introduced to describe the geometric characteristics of the microstructure. As presented in Fig. 2(a), ηf was defined by the ratio of the area with film to the whole area of a microstructure unit in xy-plane. ηc was defined by the ratio of the joint area (as shown by orange color) between the inner structure and the microstructure unit frame to the cross-section area (green and orange area) of one frame side (in yz-plane), as shown in Fig. 2(b). Thus, ηc and ηf were calculated by
To investigate the influence of the microstructure on the spatial resolution of the generated image, its modulation transfer function (MTF) was evaluated by optical knife-edge method. As shown in Fig. 3, a rectangular heat image was used to simplify the theoretical model. Under this condition, the heat had a homogenous intensity distribution in y-axis but had a step distribution along x-axis in the measurement area (as illustrated in Fig. 3 by green).
When the heat pattern was an order of magnitude larger than the size of the microstructure unit, the temperature gradient along y-axis at different x coordinate values were obtained as shown in Fig. 4, where the heat pattern size in y direction was 4 mm and the size of microstructure was 50 µm. The temperature gradient along y-axis within one microstructure (-25 µm < y < 25 µm,) can be considered as zero. Considering the small film thickness, the temperature gradient along z-axis (film thickness direction) and y-axis could be neglected, which meant that a semi-one-dimensional (s-1D) model with an adiabatic boundary condition at y=±a/2 was valid.
Semi-one-dimension meant that the model consisted of 1×N periodical microstructures. The dimension in y-axis equaled to one microstructure unit size in tens of micrometer, while the dimension in x-axis equaled to hundreds of microstructure unit size of several millimeters to ensure a sufficient thermal diffusion. When the total length L(L = Na) was one order of magnitude longer than the thermal diffusive length of the thin film chips, the adiabatic boundary condition at x=±L/2 could be applied. Thus, an s-1D heat transfer simulation model, as shown in Fig. 5, was built.
Using the s-1D model temperature distribution along x-axis at y = 0 in steady-state could be calculated with a continuous heating intensity loaded in the area of h×a. Additionally, when the heating was time-varied, the temperature variation at (0,0) point in transient-state could be simulated. Knowing the temperature of the chip, the intensity distribution of the thermal radiation along x-axis in steady-state and the thermal radiation variation at the origin in transient-state could be calculated by
and respectively. Here, ɛ is the infrared emissivity of the thin film chip and σ is the Stefan-Boltzmann constant. Based on the result of the calculation in Eq. (3), the MTF of the thin film chip could be calculated following the procedure shown in Fig. 6 [16,17]. First, the infrared radiation intensity distribution along x-axis, starting from the origin (0,0), was selected as an edge spread function (Erad (x)). Next, the line spread function (L(x)) was calculated by the spatial difference of Erad (x)(△Erad(x)/△x). Later, a spatial frequency response function (H(f)) was obtained by applying discrete Fourier transform to L(x). Through normalization and considering the positive part of the H(f), MTF was finally achieved.First, we investigated the effects of the contact area ratio (ηc) on the MTF of the chips. In this study, the contact area ratio was varied by changing w, while a, e, and l were set to be 50 µm, 3 µm, and 44 µm, respectively. The thickness of the thin film chip was 700 nm with a thermal conductivity of 1 Wm-1K-1, and infrared emissivity of 0.9 were used in the simulations. The MTF of the chips with different contact area ratios were calculated using the procedure in Fig. 6. The results are shown in Fig. 7.
The calculation results showed that the MTF decreased with the increase in the contact area ratio. When the spatial frequency with an MTF value of 0.3 was taken as spatial resolution, a linear relation between contact area ratio and spatial resolution was observed. Small ηc was helpful to improve the spatial resolution. However, due to the low-effective infrared emissivity (${\varepsilon _{eff}} = \varepsilon \cdot {\eta _f} = \varepsilon \cdot \frac{{44w + ({{50}^2} - {{44}^2})}}{{{{50}^2}}}$) with small ${\eta _c} = \frac{{w + 6}}{{50}} \times 100\%$(w was small), both maximal radiation intensity (Erad(0)) and radiation intensity contrast (Erad(0)$- $Erad(L/2)) of the chip were low, as shown in Fig. 8.
Fig. 8. Maximal radiation intensity and radiation intensity contrasts of thin film chips with different contact area ratios.
To study the effect of the different fill factors on the MTF of the thin film chips, the contact area ratio was kept at 0.18 in simulations, while the fill factor was increased by extending the total length of the inner structure (l). The MTF of chips with different fill factors are presented in Fig. 9.
The results showed that the MTF increased with the increase in the fill factor, along with the spatial resolution. When the fill factor was 0.73 the spatial resolution increased to 12.5 lp/mm. At the same time, the maximal radiation intensity and the radiation intensity contrast improved as well, as shown in Fig. 10.
Fig. 10. Maximal radiation intensity and radiation intensity contrasts of thin film chips versus fill factors.
Therefore, low contact area ratio and high-fill factor are helpful to obtain high-spatial resolution, maximal-radiation intensity, and radiation intensity contrast. Additionally, the shape of the microstructure has some influence on the spatial resolution. To improve the spatial resolution, the heat transfer should be confined in one microstructure as much as possible. Hence, microstructures with high-fill factor and long heat transfer path length are desired.
The temporal properties of the thin film chips with different microstructures were simulated by applying a time-varied heat load. The heating repetition rate was 1 Hz and the duty cycle was 10%. The normalized time dependent radiation intensity is plotted in Fig. 11, neither the contact area ratio nor the fill factor affects the thermal time response.
Based on the simulation one could see that the chip performance improved by reducing the contact area ratio and increasing the fill factor of the microstructures. To obtain low contact area ratio, narrow width of the inner pattern was required. At the same time, narrow pattern line distance had to be used to get a high-fill factor. The width of the inner pattern and the pattern line distance were limited by the fabrication ability in practice. For the chip composed of multiple thin layers, the linewidth was mainly decided by lithography accuracy. Considering the preparation accuracy of commercial lithography, a 2 µm structure linewidth could be produced. In this case, the optimized spatial resolution would be achieved by fabricating the microstructure shown in Fig. 12(a). The contact area ratio and fill factor of this microstructure were 0.12 and 0.76, respectively. Further, there were two long pattern lines connecting the inner structure to the frame. The MTF of the thin film chip is shown in Fig. 12(b), where the MTF without microstructure is plotted as well. Compared to that in the chip without microstructure, the spatial resolution of the generated infrared image was dramatically improved from 6.0 lp/mm to 15.4 lp/mm. It meant that the line pair with the width of 32.5 µm can be resolved. Since the microstructure size was 50 µm, the adjacent microstructure was distinguishable.
3. Chip fabrication and property measurements
Based on the design results, two chips with different microstructures were fabricated, which were named as ‘single-s’ and ‘double-s’. The geometric parameters of the microstructures are presented in Fig. 13. These two microstructures had nearly the same fill factor (0.74 for single-s and 0.79 for double-s) but quite different contact area ratios, which were 0.17 and 0.46 for single-s and double-s, respectively. The geometric parameters of the single-s were close to the optimized parameters by simulations.
3.1 Chip fabrication
The fabrication process of the thin film chip with in-plane microstructures is shown in Fig. 14. As the first step, PI was spin-coated as a supporting layer on a SiO2/Si substrate (the diameter was 4 inches). SiO2 was used as a sacrificial layer and would be removed in a subsequent step to form self-suspended thin film. Next, metal Cr was deposited, which served as a mask layer for PI patterning and, at the same time, as a site-selective adhesive layer for absorption/radiation layer deposition. Later, S-1813 positive photoresist was spin-rotated on the Cr metal layer. Exposed with UV-Lithography, the microstructure pattern was transferred to the photoresist. Later, the Cr layer was wet-etched to form a mask. Next, PI in the uncovered region was etched using O2-CHF3 plasma to produce in-plane microstructures on the PI layer. After etching the PI, the Cr on the microstructure frames was cleaned by overlay. The remaining Cr was used as the adhesive layer for the site-selective deposition of the absorption/radiation layer. Subsequently, the SiO2 sacrificial layer was wet-etched and the multilayer thin film peeled off from the Si substrate and supported by a PI ring. Lastly, Au black layer was sputtered in Ar + N2 atmosphere and selectively deposited on the left Cr region. The preparation of the thin film chip with in-plane microstructure was then accomplished. The front view of the fabricated chips is shown by scanning electron microscope (SEM) images in Fig. 15. It is confirmed that the microstructures were uniformly formed. Brighter parts are the microstructures with Au black layer. To reduce the heat absorption of the microstructure frame and thermal connection between adjacent microstructures, which was supposed to be helpful to improve the spatial resolution, we did not deposit the Au black on the frame of the microstructures.
As an absorption/radiation layer, Au black was deposited under 50 Pa with DC sputtering power of 90 W. The cross-sectional view of the Au black thin film is shown in Fig. 16. The film thickness was about 314 nm. The nano-porous structure ensured high optical absorption in the visible wavelength range and high emission in the infrared band.
The effective diameter of the chip was 65 mm and the chip thickness was 714 nm. The chip contained 1313×1313 microstructure units. Figure 17 shows the thin film chip.
3.2 Performance measurements
A non-contact thermography setup shown in Fig. 18 was used to measure the spatial and temporal properties. The chip was clamped into a vacuum chamber, which was used to reduce film thermal losses through heat convection and air-surface conduction. The vacuum chamber had two optical windows. The visible window was a silica window with high transmission in the wavelength range of 400 to 750 nm and with low transmission in infrared range. It allowed the low-loss passing of the optical heat and blocked the infrared radiation. The infrared window was coated for high transmission in the detected infrared band. The transmitted thermal radiation of the heated film was collected by an IR camera. To test thin film chip performances, an LCD projector (EPSON EB-C1915) and a laser, operating at 532 nm, were used. The pixel number of the projector was 1024×768, and the pixel size was 14 µm. The laser output power was modulated by its driver. In the experiments, the laser worked at 1 Hz reputation rate and the duty cycle was 40%. The laser peak power could be varied from 0 to 800 mW.
Under the steady-state condition, an optical knife-edge image in visible wavelength range was generated by the projector. The gray level of the bright half was 255 and the gray level of the dark half was 0. The output of the projector passed through the visible window and was imaged onto the thin film sample with 1:1 scale. On the film, the optical heat pattern size was 2 mm × 4 mm. The intensity uniformity of the projected pattern was 95%. The thermal radiation of the sample was detected by an IR camera (VCHD head 680 by Infratec.) with an equivalent blackbody temperature calculation function. The camera was equipped with a focal plane array that consists of 640×480 pixels with the size of 25 µm. The spectral detection range was from 7.5 to 14.0 µm, and the frame rate was 60 Hz. The noise equivalent temperature resolution of the IR camera was 30 mK. Using a lens with 1× magnification, a spatial resolution of 25 µm was obtained. The intensity data that crossed the optical knife edge were used to calculate the MTF.
Under the transient-state condition, the laser output was collimated and homogenized to a top-hat spot. The laser spot passed through an aperture with a diameter of 3 mm and was imaged onto the thin film chip by optics. When the laser output power was modulated, the thin film chip generated a time-varied infrared radiation. The infrared signal was collected by an IR camera (ImageIR@8300hp, Infratec) with a high frame rate. The frame rate of the IR camera was 255 Hz with a full frame and as high as 1200 Hz with a quarter frame. The intensity data at the laser spot center were used to evaluate the temporal properties of the chip.
Figure 19 shows the infrared image generated by the thin film chip, as well as the corresponding edge spread function in steady-state.
Fig. 19. The generated infrared image heated by the knife-edge pattern and the corresponding edge spread function.
Based on the measurement results, MTFs of the two chips with in-plane microstructures and a chip without microstructure were calculated and presented in Fig. 20. The simulation results are plotted as well.
The spatial resolution of the infrared images generated by the chips without microstructure, with double-s microstructure, and with single-s microstructure were 6.2 lp/mm, 8.0 lp/mm, and 13.2 lp/mm, respectively. Significant improvement was obtained by the chip with single-s microstructure (with lower contact area ratio). All the measurement results agreed well with the simulation predictions. The influences of the site-selective Au black deposition were weak. We assume the small thickness and the low thermal conductivity of the Au black layer to be the cause. The fluctuations in the measured MTF curves at low spatial resolution were caused by the thermal noise.
Heated by the modulated laser, the time response of the chips with double-s and single-s microstructures were measured. Figure 21(a) is the infrared image of the chip when it was heated by the laser spot. The thermal time response in the spot center are plotted in Fig. 21(b). If the decay time was defined by the time at which the radiation intensity decreased to the 1/e of its maximum, the decay time of both chips was 20 ms. No difference was observed.
In conclusion, by fabricating periodical single-s microstructures, the spatial resolution of the IR images generated by thin film chip was dramatically increased from 6.2 lp/mm to 13.2 lp/mm. While, the decay time was kept at 20 ms.
To visually show the spatial resolution, an ISO 12233 Resolution Chart was inputted to the infrared scene generator based on the thin film chip with single-s microstructure. The generated scene was collected by an infrared camera. The inputted and IR camera collected scenes are presented in Fig. 22.
Fig. 22. computer generated test scene and the corresponding IR scene generated by multi-layer thin film chip.
4. Conclusion
In this study, the performances of thin film infrared scene generation chips with in-plane microstructures were investigated. The chip consisted of substrate layer, adhesive layer, and absorption/radiation layer. Using lithography and site-selective metal black deposition, two types of microstructures were fabricated. The chip with 0.17 contact area ratio microstructures showed higher spatial resolution than the chip with 0.46 contact area ratio microstructures. Both chips presented the same thermal decay time of 20 ms. Compared to the chip without microstructure, the spatial resolution of the generated infrared image was increased from 6.2 lp/mm to 13.2 lp/mm. The experimental results coincide well with the theoretical calculations based on a semi-one-dimensional model in steady- and transient-state. Our investigation shows that the fabrication of the in-plane microstructures is an effective approach to achieve high spatial resolution as well as high frame rate. We think the spatial resolution can be further improved by using patterning technologies with high accuracy.
For comparison, the properties of mainstream infrared scene generation technologies are listed in Table 1.
Table 1. main parameters of mainstream infrared scene generation technologies
Our infrared scene generation chip showed large array size with small pixel size. Moreover, broadband infrared radiation covering mid- and far infrared was obtained. We believe the frame rate can be improved by optimizing the film materials and using active cooling.
Funding
National Natural Science Foundation of China (61704166, 61875011).
Acknowledgments
Authors are thankful to the Analysis & Testing Center and Micro-fabrication Center in Beijing Institute of Technology for providing the fabrication and characterization facility for this study.
Disclosures
The authors declare no conflicts of interest.
References
1. X. X. Xiong, F. G. Zhou, X. Z. Bai, B. D. Xue, and C. M. Sun, “Semi-automated infrared simulation on real urban scenes based on multi-view images,” Opt. Express 24(11), 11345–11375 (2016). [CrossRef]
2. X. Wu and J. Q. Zhang, “Signature simulation of infrared target by tracing multiple areal sources,” Appl. Opt. 54(13), 3842–3848 (2015). [CrossRef]
3. A. Reinov, Y. Bushlin, A. Lessin, and D. Clement, “Dew, dust, and wind influencing thermal signatures of objects,” Proc. SPIE 6941, 69410U (2008). [CrossRef]
4. S. Kim, S.-G. Sun, and K.-T. Kim, “Analysis of infrared signature variation and robust filter-based supersonic target detection,” Sci. World J. 2014, 1–17 (2014).
5. S. Lefebvre, A. Roblin, S. Varet, and G. Durand, “A methodological approach for statistical evaluation of aircraft infrared signature,” Reliab. Eng. Syst. Saf. 95(5), 484–493 (2010). [CrossRef]
6. S. Mchugh, G. Franks, and J. Laveigne, “High-temperature MIRAGE XL (LFRA) IRSP system development,” Proc. SPIE 10178, 1017809 (2017). [CrossRef]
7. R. J. Ricker, S. R. Provence, and D. T. Norton, “Broadband mid-infrared superlattice light-emitting diodes,” J. Appl. Phys. 121(18), 185701 (2017). [CrossRef]
8. R. Shi, C. Xu, R. Xu, N. Shi, Y. Yang, and L. X. Qian, “Infrared dynamic scenes generating technique based on MEMS,” Infrared Laser Eng. 45(2), 0204002 (2016). [CrossRef]
9. S. Guo, X. X. Xiong, Z. C. Liu, X.-Z. Bai, and F. Zhou, “Infrared simulation of large-scale urban scene through LOD,” Opt. Express 26(18), 23980–24002 (2018). [CrossRef]
10. A. Khalifa and H. Aly, “Near-infrared video projection system based on digital micromirror devices and digital signal processors,” Opt. Eng. 54(12), 123102 (2015). [CrossRef]
11. V. Marinove, C. H. Chi, Z. F. Tong, R. C. Liu, N. Berberova, S. H. Lin, E. Stoykova, and K. Y. Hsu, “Liquid crystal light valve operating at near infrared spectral range,” Opt. Quantum Electron. 48(4), 270 (2016). [CrossRef]
12. X. D. Lv, W. W. Wei, J. Y. Zhang, J. L. Yang, and F. H. Yang, “IR scene Generator Based on MEMS VOA Array,” AERO WEAPONRY 2016(2), 38–41 (2016).
13. F. L. Peng, D. M. Xu, H. W. Chen, and S. T. Wu, “Low voltage polymer network liquid crystal for infrared spatial light modulators,” Opt. Express 23(3), 2361–2368 (2015). [CrossRef]
14. Q. Han, J. Z. Zhang, J. Wang, and Q. Sun, “Diffraction analysis for DMD-based scene projectors in the long-wave infrared,” Appl. Opt. 55(28), 8016 (2016). [CrossRef]
15. H. D. Zhang, A. Muhammad, and J. Luo, “MWIR/LWIR filter based on Liquid–Crystal Fabry–Perot structure for tunable spectral imaging detection,” Infrared Phys. Technol. 69, 68–73 (2015). [CrossRef]
16. L. P. Ding, L. Y. Min, F. Y. Tian, and G. H. Xing, “Effect of oversample-superposition on MTF for long wave infrared camera,” J. Infrared Milli. Waves 25(1), 13–16 (2006).
17. H. Li, L. Y. Chen, G. Lei, W. D. Qu, F. Wang, and H. L. Zhang, “Analysis and implementation on canted slit-method for the test of MTF of infrared imaging systems,” Infrared Laser Eng. 33(2), 201–203 (2004).
18. R. J. Ricker, S. Provence, L. M. Murray, D. T. Norton, J. T. Olesberg, J. P. Prineas, and T. F. Boggess, “512 ( 512 array of dual-color InAs/GaSb superlattice light-emitting diodes,” Proc. SPIE 10124, 101241L (2017). [CrossRef]
19. J. R. Dupuis, D. J. Mansur, and R. Vaillancourt, “High-dynamic range DMD-based IR scene projector,” Proc. SPIE 8618, 86180R (2013). [CrossRef]
20. J. J. A. Buford, J. R. Lippert, and J. R. L. Murrer, “Record breaking high-apparent temperature capability of LCoS-based infrared scene projectors,” Proc. SPIE 7663, 76630S (2010). [CrossRef]
