Investigation of switchable radiative cooling using real-time transmittance measurement system Authors: Junbo Jung, Junjae Park, Jaehyun Lim, Joong Bae Kim Published: 2025 ABSTRACT Radiative cooling offers a passive and energy-efficient approach to thermal regulation, yet most conventional implementations lack dynamic tunability under varying environmental conditions. Recent advances in switchable radiative cooling materials have shown potential for practical applications, but their transient optical behaviors remain underexplored due to the limitations of conventional spectrophotometers. In this study, a real-time transmittance measurement system is developed that integrates a high-speed spectrometer and a photodiode for dynamic analysis of wet/dry-based switchable radiative cooling films. The system achieves a temporal resolution of 353 ms and ensures measurement reliability through simultaneous monitoring of the light-source intensity. Validation against a commercial UV–Vis spectrophotometer demonstrates high accuracy (maximum error 2.64%), and precision is confirmed through repeatability and reproducibility tests (coefficient of variation 3.57%). Using this system, the transmittance-switching behavior of porous PVDF-HFP films under various conditions is systematically investigated. The results reveal that solvent viscosity, refractive index matching, and film thickness are key factors governing response time and transmittance. Moreover, vacuum-assisted infiltration significantly enhances switching performance, achieving a wet-state transmittance exceeding 85% within 5 min. Overall, this study introduces a robust, time-resolved optical characterization platform and provides a practical methodology for evaluating the dynamic performance of switchable radiative cooling materials in realistic application scenarios. Introduction Due to global warming, the average temperature of the Earth is gradually increasing each year1, and the substantial consumption of cooling energy further accelerates this trend2. To mitigate this issue, radiative cooling technologies designed for passive cooling are gaining significant attention. Radiative cooling is a passive technique that lowers the temperature of an object below the ambient level by reflecting solar irradiance and emitting thermal energy into space, an extremely cold reservoir3–7. Its sustainable and passive nature makes it applicable in various fields, such as building design and transportation, to enhance energy efficiency in cooling systems8–14. As described above, radiative cooling technologies can significantly reduce cooling energy consumption across various fields15,16; however, their effectiveness is highly dependent on environmental factors such as solar irradiance, humidity, and cloud cover17,18. For example, in summer or hot climates where cooling demand is high, radiative cooling can effectively reduce the cooling energy consumption. However, in winter, additional energy for heating is inevitably required, while passive cooling becomes unnecessary19,20. This is because conventional radiative cooling technologies cannot dynamically adjust their optical properties (reflectance and emittance) in response to changing weather conditions21–23. It is crucial to develop radiative cooling technology that can be used universally, regardless of climate or weather conditions. Specifically, to maximize cooling performance in summer, the radiative cooling system should have higher solar-weighted reflectance and greater emittance in the sky-window spectrum. In winter, the focus should be on increasing transmittance in the solar spectrum while reducing emittance in the sky-window spectrum24,25. To address these challenges, advanced radiative cooling technologies with active tunability of optical properties based on weather conditions have gained significant attention25–27. There are several advanced radiative cooling technologies that actively switch their optical properties in response to the environment through physical mechanisms28–31. These switchable radiative cooling technologies are also being explored for integration into building envelopes, such as smart windows32,33 and building exterior surfaces34,35, where dynamically controlling solar gains and thermal radiation can reduce the energy demand for space cooling and heating. Each technology switches the optical properties of materials through a distinct mechanism. First, switchable radiative cooling technologies based on electrochromic mechanisms mainly utilize changes in an external electric field36–38. With these changes, their reflectance in the visible spectrum or emittance in the sky-window spectrum rapidly varies. Reversible metal electrodeposition is a representative electrochromic approach, implemented through the oxidation-reduction reaction of metal ions in an electrolyte. Zhao et al.39 reported a dual-mode glazing panel for switchable radiative cooling. In this study, dynamic control of incoming sunlight was achieved through the reversible electrochemical silver deposition. When a voltage was applied to activate the cooling mode, the silver film-coated glazing attained a reflectance of 89%. In contrast, in the heating mode without applied voltage, a transmittance of 70% was achieved. As a result, the system demonstrated a net cooling power of 20 − 60 W/m\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{2}$$\end{document} in cooling mode and a net heating power of 400 W/m\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{2}$$\end{document} in heating mode. Research on switchable radiative cooling technologies utilizing reversible metal electrodeposition based electrochromic approach, including this study, has demonstrated advantages such as rapid response, and precise and simple control mechanisms40,41. The mechanical approach is one of mechanisms for switchable radiative cooling technologies. Pyun et al.42 introduced a polydimethylsiloxane (PDMS) micro-nanofiber structure capable of tuning reflectance through compression and release under external force. When pressure is applied to the structure, the reflectance decreases as the pores in the PDMS are compressed. Additionally, they demonstrated that reflectance gradually change as the compression is released. Their experimental results confirmed that multi-step thermal regulation is achievable through this mechanism. Alternatively, switchable radiative cooling technologies based on mechanical approach could provide a viable solution to conventional challenges. Another approach to switching the transmittance of radiative cooling materials is to use a solvent with a refractive index similar to that of the porous medium43–45. When the solvent diffuses into the porous structure, the transmittance increases because scattering no longer occurs. Consequently, the reflectance of the radiative cooling material decreases, leading to a reduction in cooling performance. Conversely, cooling performance as well as the reflectance can be restored by removing the solvent from the porous medium. This approach allows us to actively control the optical properties of radiative cooling materials with porous structures. This approach is known as the wet/dry mechanism. Fei et al.46 proposed a porous medium for radiative cooling, synthesized using Poly(vinylidene fluoride-co-hexafluoropropene) (PVDF-HFP) and cellulose acetate, and regulated its cooling performance with ethylene glycol, which has a refractive index similar to that of PVDF. In the dry state, they achieved a reflectance of 96.6%, while in the wet state, the material exhibited 86.6% transmittance. These results demonstrate the feasibility of actively controlling radiative cooling performance through the wet/dry mechanism. Various studies have demonstrated the feasibility of switchable radiative cooling based on the aforementioned mechanisms. Furthermore, for practical implementation, it is essential to achieve not only optical switchability but also a rapid system response47–49. Zhao et al.39 and Barile et al.50 reported switchable radiative cooling technologies with response times ranging from a few milliseconds to several seconds. In these studies, the response times were typically evaluated at a single wavelength, because most conventional UV–Vis spectrophotometers employ a scanning monochromator that acquires the spectrum sequentially over several seconds or longer; when the optical properties change during the scan, the resulting spectrum no longer represents a single instantaneous state of the material. Consequently, it was difficult to assert that the switchable radiative cooling technologies could alter their transmittance across the entire visible spectrum. To overcome these limitations of conventional, scanning-based UV–Vis spectrophotometers and to provide time-resolved spectral information over the entire visible range, this study presents a dynamic transmittance measurement system specifically designed for switchable radiative cooling films. The system integrates a compact CCD-array-based spectrometer (Black-Comet, StellarNet Inc.), which acquires the full spectrum in a single shot with sub-second temporal resolution, and a parallel Si photodiode channel that continuously monitors the temporal stability of the light intensity (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_0$$\end{document}). This combination of single-shot full-spectrum acquisition, sub-second temporal resolution, and dedicated real-time source-intensity monitoring is, to the best of our knowledge, not available in conventional UV–Vis spectrophotometers or in commonly used miniature fast spectrometers. As a representative example among various mechanisms of switchable radiative cooling materials, a wet/dry-based radiative cooling material is fabricated, and its performance is evaluated using the proposed system. Its repeatability and reproducibility are validated using a commercial UV–Vis spectrophotometer and three representative liquid samples. Using the proposed system, key performance parameters of the switchable radiative cooling film, including response time, the effect of refractive index-matched solvents, and sample thickness, are systematically investigated. This work ultimately aims to establish a standardized measurement protocol and a practical evaluation platform for the advancement of switchable radiative cooling technologies. Measurement system set-up Figure 1a illustrates a schematic of the real-time transmittance measurement system proposed in this study. The system is designed to measure transmittance variations in a wet/dry-based switchable radiative cooling film using a refractive index-matched solvent. It consists of a light source unit and a detection unit. In the light source unit, white light is continuously emitted with stable intensity, while the spectrometer in the detection unit measures the intensity of transmitted light. Specifically, the light is split into two paths by a beam splitter: one directed to a spectrometer, and the other to a Si photodiode. The spectrometer continuously records the wavelength-dependent transmitted intensity for transmittance calculation, whereas the photodiode monitors the source intensity in real time. This real-time monitoring compensates for errors caused by light source fluctuations during measurement51. Figure 1b shows the actual experimental setup of the proposed system. In its current configuration, the proposed system measures direct forward transmittance under a normal-incidence, straight-through transmission geometry without an integrating sphere; therefore, the reported values correspond to directional forward transmittance rather than total hemispherical transmittance. To clarify how the proposed system overcomes the limitations of conventional instruments, a qualitative comparison is given in Table 1. Conventional UV–Vis spectrophotometers based on a scanning monochromator require several seconds or longer to sequentially scan the spectrum and thus cannot provide single-shot, full-spectrum measurements with sub-second temporal resolution. Miniaturized fast spectrometers using a linear array detector can acquire the full-spectrum in one shot with millisecond-sub-second integration times, but they generally do not offer a dedicated, user-accessible channel to monitor temporal fluctuations of the light source. In contrast, the proposed system combines single-shot full-spectrum acquisition and sub-second temporal resolution with a parallel Si photodiode that records the source intensity (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_0$$\end{document}) in real time. This combination of functionalities is unique to the proposed system and is essential for reliable time-resolved transmittance measurements of switchable radiative cooling films.Table 1Comparison between a conventional UV–Vis spectrophotometer, a miniaturized fast spectrometer, and the proposed time-resolved transmittance system.FunctionalityConventional UV–Vis spectrometerMiniaturized fast spectrometerProposed systemSingle-shot acquisition\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{a}$$\end{document}Not availableAvailableAvailableSub-second temporal resolution\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{a}$$\end{document}Not availableAvailableAvailableTemporal light sourceNot availableNot availableAvailableintensity (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_0$$\end{document}) monitoring\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{a}$$\end{document} Defined for acquiring the full-spectrum in one measurement, rather than sequentially scanning individual wavelengths Table 1Comparison between a conventional UV–Vis spectrophotometer, a miniaturized fast spectrometer, and the proposed time-resolved transmittance system.FunctionalityConventional UV–Vis spectrometerMiniaturized fast spectrometerProposed systemSingle-shot acquisition\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{a}$$\end{document}Not availableAvailableAvailableSub-second temporal resolution\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{a}$$\end{document}Not availableAvailableAvailableTemporal light sourceNot availableNot availableAvailableintensity (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_0$$\end{document}) monitoring\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{a}$$\end{document} Defined for acquiring the full-spectrum in one measurement, rather than sequentially scanning individual wavelengths Comparison between a conventional UV–Vis spectrophotometer, a miniaturized fast spectrometer, and the proposed time-resolved transmittance system. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{a}$$\end{document} Defined for acquiring the full-spectrum in one measurement, rather than sequentially scanning individual wavelengths Fig. 1(a) Schematic of real-time transmittance measurement system. (b) Image of measurement system. (a) Schematic of real-time transmittance measurement system. (b) Image of measurement system. Light source unit The light source unit of the proposed system uses a white light-emitting diode (LED) operating in the 400–700 nm wavelength range. Since the key characteristic of the wet/dry-based switchable radiative cooling film is its transmittance variation in the visible spectrum, a light source emitting visible light was selected. The employed LED (Thorlabs, MCWHL-6) has a 120\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^\circ$$\end{document} view angle, causing the light to diverge widely, which makes collimation difficult when used alone. Therefore, a collimator (Thorlabs, SM1U25-A) and an aperture were used to focus the beam into the sample while maintaining high light intensity. A 10 mm path length quartz glass cuvette was used as the sample container because the proposed switchable radiative cooling film becomes transparent when immersed in a solvent. The cuvette holder (Thorlabs, CVH100) has a 9 mm diameter input light port to ensure the incident light is accurately directed onto the sample inside the cuvette. The beam size of the generated white light was measured to be 3.8 mm using the knife-edge method (see Fig. S1). Creating a small and high-intensity collimated beam is important for high measurement accuracy, as transmittance is defined as the ratio of transmitted light to incident light. In addition, all light paths were enclosed with a black tube to block background light and reduce noise. The spectral intensity and temporal stability of the LED light source were evaluated to assess its performance (see Fig. S2). The LED exhibited high intensity in the visible range, with a pronounced peak around 450 nm, while the intensity dropped significantly outside this range. In addition, the normalized intensity remained highly stable over an 18-hour test, confirming that the light source maintains consistent optical output during long-term measurements. Detection unit As described above, a real-time transmittance measurement system that incorporates a light source intensity monitoring (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_{0}$$\end{document} monitoring) scheme to ensure reliable measurements under light source fluctuations is proposed. The detection unit consists of a spectrometer for spectral transmittance measurement and a photodiode for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_{0}$$\end{document} monitoring. A Si-based photodiode (Thorlabs SM1PD1A), chosen for its high sensitivity in the 450–650 nm range, was employed for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_{0}$$\end{document} monitoring. For spectral measurement, a Black-Comet spectrometer (StellarNet Inc., CXR-SR-50) was adopted, which uses a concave grating and micro-slit to achieve accurate spectral intensity measurements. Its fast response time further enables capturing dynamic transmittance changes of switchable radiative cooling materials. The spectral range of the spectrometer also covers the target range, and its detailed specifications are summarized in Table 2. As shown in Fig. S2a, the LED light source used in this study provides sufficient and stable intensity only within the 450–650 nm range, and the spectrometer maintains a high signal-to-noise ratio in the same range. Therefore, the analysis in this study is restricted to the 450–650 nm range to ensure reliable, high-quality transmittance measurements. With this configuration, the proposed system enables precise and reliable measurement of the dynamic transmittance of switchable radiative cooling materials.Table 2Specifications of spectrometer (StellearNet, BLACK-Comet CXR-SR-50).ParameterSpecificationsParameterSpecificationsWavelength range190–850 nmSignal to noise1000:1Dynamic range2000:1 with 6 decadesStray light0.02% at 435 nmDetector type2048 pixel CCDDetector integration1 ms–65 s Table 2Specifications of spectrometer (StellearNet, BLACK-Comet CXR-SR-50).ParameterSpecificationsParameterSpecificationsWavelength range190–850 nmSignal to noise1000:1Dynamic range2000:1 with 6 decadesStray light0.02% at 435 nmDetector type2048 pixel CCDDetector integration1 ms–65 s Specifications of spectrometer (StellearNet, BLACK-Comet CXR-SR-50). Validation unit Fig. 2Comparison of measured transmittance between commercial UV–Vis spectrophotometer and proposed system. Fig. 3The measured spectral transmittance for water and coffee samples through customized real-time measurement system, repeated 10 times for evaluation. Reproducibility: (a) Spectral transmittance in the wavelength range of 450–650 nm; (b) Mean and max/min values extracted at specific wavelengths (500 nm, 550 nm, and 600 nm). Repeatability: (c) Spectral transmittance in the wavelength range of 450–650 nm; (d) Mean and max/min values extracted at 500 nm, 550 nm, and 600 nm. Comparison of measured transmittance between commercial UV–Vis spectrophotometer and proposed system. The measured spectral transmittance for water and coffee samples through customized real-time measurement system, repeated 10 times for evaluation. Reproducibility: (a) Spectral transmittance in the wavelength range of 450–650 nm; (b) Mean and max/min values extracted at specific wavelengths (500 nm, 550 nm, and 600 nm). Repeatability: (c) Spectral transmittance in the wavelength range of 450–650 nm; (d) Mean and max/min values extracted at 500 nm, 550 nm, and 600 nm. Prior to the measurements, this study evaluates the accuracy and precision of the proposed system to assess its performance. Accuracy is defined as the difference between the measurement values obtained by the proposed system and the reference values, while precision refers to the consistency and reproducibility of repeated measurements with the same system. Since the absolute transmittance values vary depending on the sample and wavelength, the measurement error cannot be rigorously evaluated using a single reference. To address this issue, a validation process was established using three reference samples and a commercial instrument with verified accuracy and precision. Three liquid references with different spectral transmittance properties were selected, and their transmittance was first measured by the commercial instrument. These measured values were then used as reference data to evaluate the accuracy and precision of the proposed system. The commercial UV–Vis spectrophotometer (Shimadzu, SolidSpec-3700) was employed, and the selected liquid samples were water, a beverage (Powerade), and coffee. The transmittance of selected liquid samples was measured in standard forward-transmission mode using a direct detection unit accessory with a 10 mm quartz cuvette. As shown in Fig. 2, the samples exhibit distinct spectral transmittance characteristics depending on wavelength. In particular, water shows over 90% transmittance within the 450–650 nm range, whereas coffee exhibits transmittance below 20% in the same range. By employing reference samples with such distinct spectral transmittance properties, the accuracy and precision of the proposed system can be rigorously validated across a wide range of optical conditions. Figure 2 compares the transmittance of the three liquid samples measured using both the commercial UV–Vis spectrophotometer and the proposed system under identical conditions. The maximum transmittance differences were 0.55% for water at 623 nm, 2.64% for Powerade at 486 nm, and 2.44% for coffee at 650 nm. These results demonstrate that the proposed system achieves high measurement accuracy, with a maximum deviation of only 2.64% compared to the commercial spectrophotometer. In the next step, the reproducibility and repeatability were evaluated. Two samples with different characteristics of spectral transmittance, water and coffee, were selected, and each test was carried out ten times. Figure 3a and b present the results of the reproducibility test, while Fig. 3c and d show the results of the repeatability test. Reproducibility refers to the consistency of results when measurements are repeated with each sampling. In contrast, repeatability refers to the consistency of the system when the same sample is measured multiple times without re-sampling. The coefficient of variation (CV) was used as the evaluation metric for both reproducibility and repeatability. CV is defined as the ratio of the standard deviation to the mean, and a CV of less than 4% is generally considered indicative of high measurement consistency52. For the reproducibility test using water, the CV of the average transmittance across the entire spectral range was below 0.40%. Figure 3b shows that the transmittance values at 500 nm, 550 nm, and 600 nm exhibit minimal differences between the average, maximum, and minimum values. For coffee, due to its extremely low transmittance (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sim$$\end{document} 0.01%) in the short-wavelength region (500 nm), a relatively high CV was observed. However, the CV decreased significantly to 3.57% at 550 nm and 600 nm, indicating a high reproducibility of the proposed system in these regions. Figure 3c and d present the results of the repeatability test. For water, the CV remained below 0.41% across the entire spectral range. Similarly to the reproducibility test, a high CV of 17.93% was observed for coffee at 500 nm, where transmittance was extremely low (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\approx$$\end{document} 0.01%). However, the CV decreased to 2.21% and 1.02% at 550 nm and 600 nm as the reference transmittance increased. The high CV observed at 500 nm can therefore be attributed to the extremely low absolute transmittance of the sample rather than to a limitation of the proposed system. These results clearly validate the accuracy and precision of the proposed system through rigorous experiments. Switchable radiative cooling Design Fig. 4(a) Schematic of switchable radiative cooling. (b) Fabrication process of PVDF-HFP-based switchable radiative cooling film. (c) Fabricated PVDF-HFP-based porous film. (a) Schematic of switchable radiative cooling. (b) Fabrication process of PVDF-HFP-based switchable radiative cooling film. (c) Fabricated PVDF-HFP-based porous film. The transmittance of a wet/dry switchable radiative cooling film was analyzed using the proposed real-time transmittance measurement system. The film has a porous structure based on PVDF-HFP. Figure 4a illustrates the switching mechanism of the PVDF-HFP-based radiative cooling film. In the dry state, strong scattering occurs in the solar spectrum because of the refractive index mismatch between the polymer matrix and the air-filled pores. This results in high reflectance in the solar spectrum, thereby reducing the heating effect caused by solar irradiance and enabling effective cooling performance. In contrast, when a solvent with a refractive index similar to that of PVDF-HFP diffuses into the pores, the refractive index contrast is significantly reduced, leading to decreased scattering and increased transmittance. This switchable functionality can be utilized effectively in environments with large temperature fluctuations or during winter when heating is required. To investigate the practical applicability of radiative cooling, a PVDF-HFP-based porous film was fabricated and its transmittance change was quantitatively analyzed depending on the presence or absence of a refractive index-matched solvent. Fabrication The PVDF-HFP-based radiative cooling film was fabricated through a four-step process, as illustrated in Fig. 4b. The fabrication method and mixing ratios were adopted from previous studies53,54. First, PVDF-HFP pellets (Sigma-Aldrich LLC.) and acetone were mixed at a weight ratio of 1:8. The mixture was stirred at 300 rpm for 60 minutes, while heated to 40\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document} using a hotplate with a magnetic stirrer to ensure homogeneous dissolution. In the subsequent phase separation step for the formation of a porous structure, water was introduced into the solution—added at the same weight as PVDF-HFP. Water is highly miscible with acetone but immiscible with PVDF-HFP, thereby inducing phase separation. After the addition of water, a second stirring step was performed under the same conditions. Finally, the resulting mixture was cast onto a glass substrate. As the acetone and water evaporated, nano/micro-scale pores formed within the PVDF-HFP matrix. According to previous studies, porous PVDF-HFP films exhibit mainly interconnected hierarchical porous structures with mean nanopore and micropore sizes of approximately 200 nm and 5.5 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}, respectively53,55. Since the mixing ratio and treatment conditions in this study closely follow these reports, the fabricated PVDF-HFP film is expected to exhibit a similar micro/nanoporous morphology. Figure 4c shows the dried PVDF-HFP-based switchable radiative cooling film. The dry-state film exhibits high reflectance in the solar spectrum due to strong light scattering and appears white in color. To quantify these optical properties, the spectral reflectance and transmittance of dried PVDF-HFP films were measured in the range of 0.3–15 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document} (see Fig. S3). Films with three different thicknesses show high reflectance (> 77.8%) in the solar spectrum and transmittance lower than 11.3%. In particular, for the 336 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}-thick sample, the reflectance and transmittance are 90.3% and 6.0%, respectively, confirming that increased thickness reduces transmittance and enhances reflection. In addition, the dry-state film exhibits high emissivity (84.5%) within the 8–13 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document} sky-window spectrum, which is essential for excellent radiative cooling performance. Results and discussion Preliminary test of switching behavior In this study, a real-time transmittance measurement system was developed to quantitatively evaluate the performance of a switchable radiative cooling film. Among various switchable radiative cooling strategies, a refractive index-matched solvent approach was employed to control the film transmittance. Detailed information on the solvents is provided in Table 3. The refractive index of the PVDF-HFP matrix used in this study is 1.407 at 589 nm. The refractive indices of the selected dimethyl silicone fluids (silicone oils) are 1.382 for SF1000N-1 cSt and 1.399 for SF1000N-10 cSt, indicating that the 10 cSt silicone oil more closely matches that of PVDF-HFP. The optical transparency of the silicone oils in the 300–2500 nm range was confirmed using a commercial UV–Vis spectrophotometer, showing an average transmittance above 90% in the visible spectrum (see Fig. S4).Table 3Summary of silicone oil characteristics.Product nameViscosityRefractive indexVolatile contentBoiling point25\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document} [cSt]25\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document}150\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document}, 3 hour [%][\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{\circ }\textrm{C}$$\end{document}]SF1000N-1 cSt11.382−153SF1000N-10 cSt101.399< 30− Table 3Summary of silicone oil characteristics.Product nameViscosityRefractive indexVolatile contentBoiling point25\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document} [cSt]25\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document}150\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document}, 3 hour [%][\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{\circ }\textrm{C}$$\end{document}]SF1000N-1 cSt11.382−153SF1000N-10 cSt101.399< 30− Summary of silicone oil characteristics. As a preliminary test, the switching behavior of the radiative cooling film was visually observed. Figure 5a and b show the time-dependent optical changes of the same film immersed in silicone oils with different viscosities. A distinct difference in response time was observed between the two cases. In 1 cSt silicone oil (Fig. 5a), a rapid transition began within 5 s, and the film became uniformly transparent after approximately 20 s. In contrast, in 10 cSt silicone oil (Fig. 5b), the transition occurred more slowly, reaching full transparency only after about 10 min. In both cases, the transition proceeded from the edges toward the center of the film.Fig. 5Photograph of real-time changes depending on (a) 1 cSt silicone oil and (b) 10 cSt silicone oil. Real-time average transmittance changes using (c) 1 cSt silicone oil (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t =111 ~\upmu$$\end{document}m) and (d) 10 cSt silicone oil (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t = 101 ~\upmu$$\end{document}m). Photograph of real-time changes depending on (a) 1 cSt silicone oil and (b) 10 cSt silicone oil. Real-time average transmittance changes using (c) 1 cSt silicone oil (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t =111 ~\upmu$$\end{document}m) and (d) 10 cSt silicone oil (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t = 101 ~\upmu$$\end{document}m). For the quantitative analysis, the real-time transmittance of the samples was measured using the developed system, as shown in Fig. 5c and d. The initial average transmittance of the dry-state PVDF-HFP film was as low as 0.05%, indicating an opaque state. When immersed in 1 cSt silicone oil, the average transmittance increased to 35.03% within 111 s and converged to 35.78% after 210 s. In contrast, as shown in Fig. 5d, the transmittance change in 10 cSt silicone oil was only 1.35% during the same 111 s, eventually increasing to 54.08% after approximately one hour. The difference in transmittance response time can be explained by Darcy’s law (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$Q=-kA \Delta P/(\mu L)$$\end{document}). According to the law, under identical permeability (k), pressure gradient (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta P$$\end{document}), flow length (L), and cross-sectional area (A), the volumetric flow rate (Q) through a porous medium is inversely proportional to the fluid’s dynamic viscosity (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mu$$\end{document}). Thus, higher viscosity results in slower solvent transport. In this experiment, the time to reach 35.78% transmittance was 111 s for the 1 cSt oil and approximately 1500 s for the 10 cSt oil, qualitatively consistent with the 10-fold viscosity difference. Although not exactly proportional, this result suggests that the switching behavior of the film can be reasonably explained by Darcy’s law. In addition to viscosity, the differences in final transmittance can be attributed to the solvent’s refractive index. While not visually apparent in Fig. 5a and b, the proposed measurement system revealed that the film immersed in 10 cSt silicone oil exhibited higher final transmittance. The difference in average final transmittance between the two cases was approximately 20%. This result is consistent with the fact that the refractive index of 10 cSt silicone oil is 0.017 higher than that of 1 cSt oil and closer to that of PVDF-HFP (see Table 3). Overall, the proposed switchable radiative cooling film operated successfully using a refractive index-matched solvent. The experimental results demonstrate that the parameters included in Darcy’s law-such as permeability, dynamic viscosity, and cross-sectional area-together with the solvent’s refractive index, strongly influence both the transmittance response time and the final transmittance. Cycling test For the cycling test, all films were fabricated with a thickness (t) of 60 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}. Cycling tests were repeated five times, with transmittance monitored over a 10 min period for each cycle. In order to make wet-state, the dry-state PVDF-HFP film was immersed in a cuvette filled with solvent, and its transmittance was recorded for 10 min. The film was then retrieved and dried at 100\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document} for 10 min to restore its dry state. Figure 6a and b present the cycling results for silicone oils with viscosities of 1 cSt and 10 cSt, respectively.Fig. 65 repeated wet/dry states cycling test based on (a) 1 cSt silicone oil and (b) 10 cSt silicone oil. 5 repeated wet/dry states cycling test based on (a) 1 cSt silicone oil and (b) 10 cSt silicone oil. When 1 cSt silicone oil was employed, the average transmittance increased from 0.20% (dry state) to 59.29% (wet state) in the first cycle, demonstrating a faster and higher switching performance than in the previous refractive index matching test. In contrast, with 10 cSt silicone oil, the transmittance increased from 0.032% to 58.79%. The time required to reach an average transmittance of 50% was 243 s for the 1 cSt silicone oil and 459 s for the 10 cSt silicone oil, indicating that the higher viscosity delayed switching by approximately 1.8 times. While this delay can be attributed mainly to viscosity, the relationship is not strictly proportional due to variations in pore morphology, size, and porosity. Prior studies have reported that non-uniform pore structures in PVDF-HFP films can arise from incomplete mixing of acetone and water during phase separation, since the pore morphology is highly sensitive to solvent-nonsolvent mixing conditions53,56,57. From the second cycle onward, a different behavior was observed. For the 1 cSt silicone oil, the saturated average transmittance decreased from 59.29 to 19.5%, a drop of 39.79%. In contrast, for the 10 cSt silicone oil, the average transmittance after drying increased by 12.08%, and rapid switching of 57.02% occurred within just 1.5 s after immersion. These trends remained consistent in the subsequent cycles. This behavior-such as altered wet/dry transmittance values and the accelerated switching in the 10 cSt condition-can be attributed to the evaporation characteristics of the solvents. As shown in Table 3, 1 cSt silicone oil has a boiling point of 153\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document}, whereas 10 cSt silicone oil shows only \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sim$$\end{document} 30% weight loss after 3 h at 150\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document}. As shown in Fig. S5, an evaporation test conducted at 100\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document} revealed that 12.1% of the 1 cSt silicone oil evaporated within 10 min, while only 0.5% of the 10 cSt silicone oil was lost. This suggests that the 1 cSt silicone oil was nearly completely removed during drying, allowing full recovery of the dry state, while the 10 cSt silicone oil remained partially trapped within the film, hindering complete drying. Moreover, the excessive heating in the 1 cSt silicone oil case likely caused thermal shrinkage of the PVDF-HFP matrix, reducing pore size and hindering solvent diffusion. In contrast, in the 10 cSt silicone oil case, the residual solvent increased the dry-state transmittance and preserved a less viscous surface environment, both of which contributed to faster switching. The film appearance during cycling is shown in Fig. S6a. To isolate the effect of thermal shrinkage, an additional cycling test was conducted at room temperature (28\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\,^{\circ }\textrm{C}$$\end{document}) using 1 cSt silicone oil. As shown in Fig. S6b, no reduction in average transmittance was observed over repeated cycles, supporting the conclusion that the performance degradation observed previously was primarily due to thermal shrinkage. These findings demonstrate that the transmittance and response time of switchable radiative cooling films are governed not only by pore structure and solvent thermodynamics but also significantly by the drying temperature. Thickness effect Finally, the average transmittance and response time under wet/dry conditions for films of different thicknesses were analyzed using the proposed real-time transmittance measurement system. 10 cSt silicone oil, selected for its relatively high transmittance, was used as the solvent. PVDF-HFP films with identical compositions but different thicknesses (75 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document} and 351 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}) were fabricated, and the measurement results are presented in Fig. 7a. In the dry state, the transmittance of the two films was as low as 0.19% and 0.05%, respectively, both within the measurement error range. The time required to reach a steady wet-state transmittance was 1 h for the thin film (75 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}) and 23 h for the thick film (351 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}). The final steady wet-state average transmittance was 63.88% for the thin film and 57.79% for the thick film, indicating that transmittance decreases with increasing thickness, even after full solvent diffusion. This reduction is attributed to enhanced light scattering within the film due to a longer optical path length, as reported in previous studies58. Additionally, slower solvent penetration into the deeper porous structure of the thicker film is considered a contributing factor. To improve the practical applicability of switchable radiative cooling films, vacuum processing was introduced during the dry-to-wet transition. A vacuum chamber capable of generating sub-atmospheric pressure was employed to evacuate air from the porous structure of the PVDF-HFP films and accelerate solvent diffusion. Films with thicknesses of 66 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}, 157 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}, 202 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}, 251 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}, and 355 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document} were prepared, and a vacuum processing was uniformly applied during the wetting stage for 5 min at a pressure below 0.1 MPa. Figure 7b shows the average transmittance of each film in the dry state and after 5 min of vacuum-assisted wetting. As shown by the blue line in Fig. 7b, the dry-state transmittance of all films remained below 0.2%, indicating negligible light transmission. In contrast, the red line represents the wet-state transmittance. Unlike in the previous refractive index matching and cycling tests, vacuum processing enabled the thinnest film to achieve an average transmittance of 85.01% within 5 min under the 10 cSt silicone oil condition-a substantial improvement over the previous result, where the 75 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}film required 1 h to reach 63.88%. Furthermore, wet-state transmittance decreased with increasing film thickness; even the thickest film (355 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}) reached 33.98% within 5 min. Although extended vacuum processing could yield higher transmittance, the measurement time was limited to 5 min to reflect the rapid switching requirements of practical applications. Overall, this study quantitatively analyzed the time-dependent optical performance of switchable radiative cooling films using the proposed real-time transmittance measurement system-a capability not achievable with conventional UV–Vis spectrophotometers. The real-time average transmittance of PVDF-HFP films was evaluated as a function of film thickness, processing conditions, and solvent type. It was confirmed that vacuum processing enables practical response speeds for films thinner than 100 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}. Moreover, this case study demonstrates that the proposed measurement system is broadly applicable to research fields involving dynamic changes in optical properties.Fig. 7(a) Average transmittance and response time of PVDF-HFP samples (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t=75~\text {and}~351$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}) in dry and wet states. (b) Average transmittance of PVDF-HFP samples with different thickness using vacuum chamber. (a) Average transmittance and response time of PVDF-HFP samples (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t=75~\text {and}~351$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu \,\hbox {m}$$\end{document}) in dry and wet states. (b) Average transmittance of PVDF-HFP samples with different thickness using vacuum chamber. Conclusions This study proposed a real-time transmittance measurement system capable of characterizing the time-dependent optical behavior of switchable radiative cooling materials in the visible spectrum. The system, composed of a high-speed spectrometer and a photodiode, demonstrated rapid acquisition capabilities with high precision and reliability, comparable to those of commercial UV–Vis spectrophotometers. Specifically, the maximum measurement error was 2.64%, and the coefficient of variation for repeatability and reproducibility was measured at 3.57%. Using this system, the transmittance-switching behavior of PVDF-HFP films with porous structures was investigated under wet/dry conditions. These films exhibited dynamic transmittance changes by absorbing silicone oil with a refractive index similar to that of the PVDF-HFP matrix. Experiments with 1 cSt silicone oil showed a transmittance increase from 0.05 to 35.8% within 210 s, while 10 cSt silicone oil achieved 54.08% over 1 h. Cycling tests revealed a gradual decrease in switching performance with 1 cSt oil, whereas 10 cSt oil showed an increasing trend, reaching up to 84.44% after repeated cycles. Film thickness was found to be a critical parameter: thinner films enabled faster transitions and higher transmittance. Furthermore, vacuum processing was introduced to accelerate the transition and enhance performance, achieving a maximum wet-state transmittance of 85.01% within 5 min. In summary, the proposed real-time transmittance measurement system enabled quantitative evaluation of key factors influencing switchable radiative cooling performance, including solvent viscosity, refractive index matching, cycling behavior, and film thickness. The results demonstrate the system’s potential as a valuable platform for dynamic optical analysis and suggest its applicability in various industrial applications, such as smart windows and advanced thermal management systems. In addition, access to full-spectrum, time-resolved transmittance data enables wavelength-resolved switching times to be extracted and linked to material and process parameters, thereby advancing the understanding and design of switchable radiative cooling materials beyond simple comparisons of initial and final states. Although the present implementation has been experimentally demonstrated only for wet/dry-based switchable radiative cooling films using a cuvette-type sample holder, the same time-resolved transmittance setup can, in principle, be extended to other mechanisms of switchable radiative cooling systems by replacing the sample holder and applying appropriate driving conditions for each switching mechanism, which will be explored in future work. Supplementary Information Supplementary Information. Supplementary Information.