Diarylethene Photoswitches and 3D Printing to Fabricate Rewearable Colorimetric UV Sensors for Sun Protection

Despite education campaigns linking sun overexposure and skin cancer, it remains one of the leading preventable cancer diagnoses. Skin cancer risk is correlated with overexposure to UV light in sunlight and can be prevented by avoiding exposure. While sun protection can be achieved using sunscreen and clothing, people must be made aware of their risk to facilitate behavior change. Herein, new rewearable UV sensors which overcome the single‐use limitations of other products are presented. These sensors utilize diarylethene photoswitches, which develop a colored appearance upon exposure to UV and are reset to colorless by green light (<10 min). These photoswitches are incorporated into a range of materials, enabling the use of advanced manufacturing to develop highly desirable consumer products. 3D stereolithographic printing is used to prototype sensors, with complex geometries and appealing aesthetics, that can be worn by users. The UV sensitivity of these devices is tuned by incorporating chromophores, meeting the needs of diverse skin types. The colorimetric response allows for direct visual feedback to the user, or quantification using photography, allowing for dosimetry of UV exposure. These new reusable devices aim to reduce people's exposure to UV, while reducing the waste generated by single‐use devices.


Introduction
It is often what we cannot see or feel that harms us the most. This is true for UV radiation, which can negatively affect our health, with limited ways for us to sense it until after either photoelectric or photochromic, based on their mechanism of signal generation. While all devices can detect UV radiation, they trade off in a range of other critical factors that affects their success as wearable sensors. It was critically identified that approaches that do not require power, can discriminate between UV-A and UV-B, provide a sensitive and cumulative dose-response (dosimetry), and an instantaneous read out without the need for additional device intervention are likely to lead to the improved success of wearable devices. [9] As such, simple technologies based on photochromic materials are exciting for further innovation in this field.
Photochromic devices rely on a chromic species that undergoes a visual color change in response to light. Many of the devices reported previously rely on either a change in pH or redox potential, mediated by a secondary colorless photoactive molecule. A sophisticated example has combined these approaches, using the oxidation of methylene blue into leucomethylene blue, as well as the deprotonation of neutral red, in a matrix of tetramethylenediamine acrylic acid. [11] While these devices worked on a negative response (from colored to colorless), positive dose-responses are potentially more beneficial, as they are easier to read and interpret by eye. Inorganic nanoparticles can also be used in photooxidation-based sensors, and a recent example developed a positive response ink that could be printed onto paper-based devices. [12] These devices relied on the photooxidation of phosphomolybdic acid in the presence of lactic acid. They were highly sensitive, selective for UV-B light, and could be personalized for different skin tones using neutral density filters. The challenge for these photochromic devices is a lack reversibility of the chemistry, therefore making them single-use devices. In the current societal climate for reducing waste, reusable and rewearable devices are highly desirable. [13] Photoswitches are another class of photochromic materials that can be used to develop UV sensors. Photoswitches undergo an intramolecular rearrangement (cyclization or isomerization) upon light irradiation, which can then be reverted either thermally (T-type), or by irradiation with a light of a different wavelength (P-type). [14,15] This reversibility of the photochemistry makes them ideal candidates as reusable UV sensors. SPOT-MYUV (Dig It Apparel Inc., Canada) have used spiropyran photoswitches to develop sunscreen reminder stickers. [16] The stickers are attached to the skin, and when exposed to UV radiation become purple. As T-type photoswitches, when sunscreen is applied to the body and the sticker, the UV light is blocked, and they quickly discolor. When protection from the sunscreen is no longer effective, the change in color of the sticker alerts the user to apply more sunscreen. While the technology that underpins these stickers is reusable (multiple color changes possible per wear), as stickers the devices are not rewearable. Also, the devices only act as reminders to seek sun protection or apply sunscreen, and do not provide wearers with their accumulated dose or risk for the day.
There is a clear gap in the rewearable UV technology space for rewearable dosimeters that can modify user behavior, as well as inform on dose and risk. Another challenge is the diversity of skin tones in the population, which are often categorized using the Fitzpatrick scale. [17] These different skin categories have different sensitivities to UV light, often characterized by a minimum erythema dose (MED). [17] Current technologies favor low dose-responses that suit fair skin tones (typically class II), to prevent sunburn. This leads to disadvantage for people with darker skin tones (class III-VI), which are also at risk of sunburn, or at risk of too low an UV dose for effective vitamin D metabolism. [8] Finally, there is a critical need to address behavior change in sun seekers, particularly adolescents and young adults, where increased exposure at these ages has been associated with greatly increased risks of melanoma later in life. [3,18] Current technologies based on stickers, paper, wristbands, and complicated electronic devices do not meet the design demands for end users. [13] In this project, we have developed highly desirable, reusable, and simple UV sensors for sun protection (Figure 2). This was Adv. Mater. Technol. 2023, 8, 2201918 Figure 1. The International Commission on Non-Ionizing Radiation Protection (ICNIRP) action spectrum for the risk to human skin and eyes from UV light (---, left axis) compared to the terrestrial solar spectrum (-, right axis). [6] Figure 2. Overview of the development of rewearable UV sensors based on photoswitches, from molecular design to device prototyping, and performance evaluation. www.advmattechnol.de achieved by using diarylethene photoswitches as highly sensitive photochromic materials, that go from colorless to dark pink on exposure to UV radiation, that can be restored to their original state by green light. The ability to choose when to reset the devices, rather than a spontaneously reversible process, allows for the measurement of UV dose across a day and the potential to communicate about the level of risk to the wearer. The diarylethene photoswitches can be simply incorporated into a range of materials for device fabrication, including silicone or resins for 3D stereolithographic (SLA) printing, while maintaining their sensitive UV response. The tuning of the sensitivity and selectivity for UV-A and UV-B will be described, as well as simple quantitation through photography and image analysis. The production of desirable designs using advanced manufacturing techniques demonstrates the applicability of this approach for modern consumer markets.

Synthesis and Evaluation of Diarylethene Photoswitch as an UV Sensor
This project aimed to develop a photoswitch to be used as an UV dosimeter inside plastic materials, providing smart wearable sun sensors. Therefore, the desired photoswitch must fulfill the following requirements: i) absorbance in UV-A and UV-B region (280-400 nm), ii) reversible isomerization in the solid state, iii) visually obvious color change during irradiation, and iv) resistance to fatigue over repeated use. The diarylethenes (DAE) were selected as the photoswitch of choice, for their capacity to meet all these requirements. [19,20] The photoswitch consists of two heterocyclic rings connected via a bridging unit like cyclopentene or maleimide. [21,22] The cyclopentane bridging structure has become the predominant DAE, owing to the cheaper and facile synthesis at scale. Therefore DAE (5) was chosen as a suitable candidate and was prepared according to literature procedures (Scheme 1). [23] Briefly, the synthesis of 5 was achieved, by starting with commercially available 2-methylthiophene (1), which was chlorinated with N-chlorosuccinimide and distilled under reduced pressure to achieve compound 2 (52%). In the next step, a Friedel-Crafts acylation with glutaryl chloride gave product 3 in a low yield (15%), due to an undesired side reaction with glutaryl chloride and the methyl group of the thiophene. Despite the low yield, the subsequent McMurry reaction of 3 with TiCl 4 and zinc provided the diarylethene precursor 4 (35%). The final DAE was achieved via a Suzuki-Miyaura cross-coupling reaction providing 5 in a high yield (81%). Overall, this synthetic strategy has been used to produce 5 up to a 1 g scale , suitable for prototyping a range of UV dosimeter designs.
To validate that diphenyl diarylethene can act as a sensitive probe for UV radiation, at doses relevant for sun protection, a series of solution-based experiments were performed. A stock solution of 5 in toluene was prepared, and aliquots were irradiated using UV-A and UV-B fluorescent lamps in a photoreactor for increasing UV doses. Toluene was selected as the solvent to filter any residual UV-C emission from the lamps, which is not relevant for sun protection in terrestrial applications. The absorbance of each solution was measured at 530 nm using an UV-vis spectrometer and plotted against the dose delivered and the corresponding MED for each Fitzpatrick skin type (Figure 3). Comparing UV-A to UV-B dose-response curves, it is clear that the photoswitch is more sensitive to UV-B light (1.6-fold greater at 2100 J m −2 ). This is beneficial, as UV-B is considered the major risk for sun-related diseases such as skin cancer. [4,5] The sensitivity of human skin for UV-B to UV-A is ≈1000-fold greater (Figure 1), [6,17] so despite the enhanced sensitivity for UV-B, an improvement in selectivity is still required.
In previous reports that have aimed to tune the spectral sensitivity of wearable UV sensors, neutral density films and other bandpass filters are assembled on top of the UV sensitive material. [12,24] This approach requires assembly of multiple components, complicating fabrication processes. In this project, we aimed for fully 3D printable devices, so it was hypothesized that increased selectivity for UV-B could be achieved by including a chromophore in the detector matrix that absorbs strongly in the UV-A region, with a much lower extinction coefficient in the UV-B region. In a first attempt anthracene was used, demonstrating that the response to UV-A could be switched off, over the typical MED dose range for UV-B of the six skin types ( Figure 3). While the UV-B sensitivity was also decreased, it still showed a dose-response and a greater sensitivity compared to UV-A (9.5-fold greater at 2000 J m −2 ). Anthracene is not an ideal candidate since it has highly variable extinction coefficients across the UV-A region ( Figure S1, Supporting Information). To improve the selectivity of the sensors, a range of chromophores were considered as filters for www.advmattechnol.de the devices, using published UV-vis spectral data. [25] 7-Diethylamino-4-methylcoumarin was selected as a desirable candidate for its broad absorption across the UV-A spectrum, low absorption in UV-B region, and higher extinction coefficient compared to anthracene ( Figure S1, Supporting Information).
While UV-B is generally considered as the greater risk for disease through direct DNA damage, there is evidence to support UV-A may play a role through the activation of ROS. [4,5] Therefore, being able to tune the sensitivity of devices to UV-A and UV-B is highly beneficial, as our understanding of sun-related diseases evolves. The effect of varying ratios of coumarin on the specificity of diphenyl diarylethene for UV-B light was investigated (Figure 4). While the inclusion of coumarin decreased the sensitivity for both UV-A and UV-B light, the effect was clearly more significant on UV-A at both low and high doses. By comparing the ratio of the absorbance at 540 nm for samples irradiated with UV-B to UV-A, the specificity for UV-B light can be maximized between 10 and 50 equivalents of coumarin to photoswitch (up to fivefold greater). From this experiment, the optimal ratio is ten equivalents, to maximize the specificity for UV-B, while maintaining sensitivity in terms of colorimetric response.
The materials developed in this project aimed to take a lowtech approach, providing an instantaneous read out of dose by eye based on color change. Despite this goal, the ability to quantify this color change may be of benefit to some users, for public health research, or for UV-protection in occupational health scenarios. Therefore, photographic images of the irradiated solutions were taken and analyzed using image processing software and compared to the spectrophotometric measurements. Both the RGB and HSB color space were investigated for their ability to quantify photoswitching. [26] For the solutions, both the green channel of RGB, and the saturation channel of HSB, showed a strong correlation with the spectrophotometric absorbance at 500 nm ( Figure S3, Supporting Information). Other channels did not demonstrate dose-response or had a lower dynamic range. This result provides a simple approach for quantification of solid and 3D-printed devices developed in this project, as well as the potential to use smartphones to quantify devices in use.

Flexible Silicone Devices
The first attempt to develop a solid 3D sensor utilized silicone resin as the structural material. Silicone was chosen as it is cheap, safe, and has a high optical transmittance, and so is often used to create cheap UV wearable devices. It also allowed for simple casting of the diphenyl diarylethene photoswitch into the matrix, and production of thin flat sensors for simple UV dose testing. Preparation of the sensors involved careful grinding of the recrystallized diarylethene to a very fine powder, dispersing into the resin component of the silicone kit, before blending with the catalyst component. The resin could then be cast into a variety of molds to produce sensors of varying shapes and sizes. Thin, flat sensors were cast from a square dish for evaluation of the UV dose sensitivity.
Silicone UV sensors were cast with 0.02 wt% diphenyl diarylethene to test for dose sensitivity and reversibility. The sensors were irradiated with UV-B light (25 W m −2 ) in a photoreactor, with strips of aluminum foil being progressively removed over time, providing different dose exposures along the strips. The dose-dependent color change was clearly visible to the naked eye (Figure 5a). The sensors were photographed using a smartphone, or scanned using a flatbed scanner, to quantify the color change relative to UV-B dose (Figure 5b). The mean pixel intensity for the green channel of RGB images, or hue and saturation of HSB images, demonstrated good correlation with UV-B dose, with the dynamic range of the 0.02 wt% sensors covering the MED range for all six Fitzpatrick skin types. The green channel of both photographs and scanned images provided the most sensitive detection, with no significant difference between photography or scanning, so RGB photography was used for the remainder of the study.
With the goal of developing reusable devices, the reversibility of the DAE-silicone sensors was tested using a single   www.advmattechnol.de low powered green light-emitting diode (LED, 1 W, 520 nm). The sensors were irradiated and photographed at various time intervals for up to 2 h, with complete discoloration achieved in 60 min (Figure 5c). Inspired by the storage boxes and stands for Bluetooth wearable devices, which charge the devices while stored, a prototype storage box fit with 4 × 1 W LEDs was built, to enable the safe and simple resetting of devices in the home ( Figure S4, Supporting Information). Increasing the number of LEDs and therefore intensity of light, speeds up the reversal process, as seen below with the 3D-printed beads (Section 2.2.2).
To reduce exposure to UV radiation when in the sun, many people will demonstrate primary prevention behaviors, such as seeking shade, wearing protective clothing, and applying sunscreen. [3] Wearable UV sensors should consider these behaviors when providing assessments of risk. To demonstrate that silicone-DAE sensors are compatible with sunscreen, a sample was coated with the recommended dose of 2 mg cm −2 of SPF50+ sunscreen and exposed to 375 J m −2 of UV-B light (MED for Fitzpatrick skin type III 300 J m −2 ). It is clear from the images that the sample coated with sunscreen shows no significant change in color, compared to the uncoated sample which has turned the distinctive pink color of the ring closed DAE ( Figure S5, Supporting Information). This demonstrates that the DAE sensors of this project can perform comparably to other sunscreen compatible technologies in the market.

3D-Printed Wearable and Reusable UV Sensors
While a range of unpowered UV sensors is available in the market, these are generally limited to simple stickers and wristbands. While the devices can sensitively detect UV and convey some message of associated risk to the wearer, they have had contradictory results on improving sun safe behaviors. [27,28] Factors that could be reducing the efficacy of these devices are the desirability and everyday wearability of their designs. Poor aesthetics, limited customization, complexity, and incompatibility with personal identity are found to be under-addressed factors for disengagement with wearable technology. [29,30] Attempting to make UV sensors more appealing, interactive, and applicable to daily life, this project aimed to develop 3D-printed devices which can be fabricated into nearly any shape, allowing for production of diverse and customizable wearable devices, such as jewelry, buckles, and sunglass frames. Stereolithographic printing was selected owing to its high print quality, capability to produce complex and varied geometries, and ability to easily mix the DAE photoswitch into the liquid photopolymer printing resins. A clear resin was prepared with carefully ground diphenyl diarylethene (0.03 wt%). Exemplary prints for a range of devices are shown in Figure 6 before and after exposure to natural sunlight (90 s), as well as resetting with green light (90 min, 4 × 1 W LED). Clearly from the first photograph, the 405 nm laser used in the SLA printer does not activate the DAE photoswitch during fabrication. A clear and distinctive color change is visible upon exposure to natural sunlight, even at the low loading of chromophore in the resin. Once again, the DAE photoswitch is easily reset in the 3D resin, allowing for devices to be reused.
To confirm that the 3D-printed DAE chromophores could be used as reusable UV sensors, the dose-response and stability   www.advmattechnol.de over multiple cycles were tested. A set of small 3D-printed beads, suitable for jewellery or decoration, was used to demonstrate that sensors with complex shapes could still be quantified ( Figure S6, Supporting Information). First, the dose-response to UV-B light was demonstrated, with the beads showing a strong colorimetric response, which was quantified using photographic analysis (Figure 7a). Three beads were then exposed to a series of cycles of UV-B light at a high dose equivalent to the MED for Fitzpatrick skin type VI (1000 J m −2 ) and resetting with the prototype LED boxes for 10 min (4 × 1 W LED). Again, photographic analysis was used to quantify dosimeter performance, with the green pixel intensity min-max normalized using irradiated and unirradiated bead samples for consistency across every image ( Figure S7, Supporting Information). It is clear over the ten cycles that the three beads performed consistently, providing a normalized mean pixel intensity of 90 ± 17 AU for 1000 J m −2 of UV-B, and then returning to the same colorless state after resetting with the LEDs (−2 ± 18 AU). This demonstrated stability of the DAE in the 3D-printed resin, over multiple cycles, proves that these devices offer a reusable and therefore potentially more sustainable solution than other commercially available, low-tech products.

Conclusions
This project demonstrates that diarylethene photoswitches can be developed into wearable UV radiation sensors, able to provide wearers with instantaneous feedback on their relative dose and risk from sunlight. The photoswitches and the UV sensitive materials in this project have been demonstrated to be highly sensitive to UV light, easily reset using green light and stable over multiple cycles, allowing for one device to be reused many times. It is possible to tune the sensitivity between UV-A and UV-B light, allowing for better determination of diseaserelated risk. The materials have also been demonstrated to have dosimetric capability, allowing for monitoring of the level of UV exposure for the wearer using simple photographic analysis. Future work aims to embed the photographic analysis into an automated program, compatible for use with smartphone technology.
Using the synthesized diarylethene photoswitches, we have produced prototype UV sensors using 3D printing. This demon strates the compatibility of these molecules with modern manu facturing processes, allowing for straightforward production of nearly limitless designs. The intention of using 3D printing is to enable the production of highly desirable products and wearables, encouraging users to integrate these sensors into their everyday life. Hopefully more seamless integration into the daily lives of users offers the possibility to improve sun safe behaviors and reduce the risk of UV-related diseases.

Experimental Section
Materials: Unless otherwise stated, all solvents and reagents were purchased from Sigma-Aldrich (Castle Hill, NSW, Australia) or Thermo Fisher Scientific (Scoresby, Australia) at analytical grade (or higher) and used without further purification. Vario 15 A/B silicone rubber kit was purchased from Barnes (Brisbane, Australia). Formlabs Clear V4 resin was purchased from Thinglab (Yarraville, Australia).
Chemical Characterization: 1   www.advmattechnol.de equipped with an HESI II probe. The instrument was calibrated in the m/z range 74-1822 using premixed calibration solutions (Thermo Scientific). A constant spray voltage of 3.5 kV, a dimensionless sheath gas, and a dimensionless auxiliary gas flow rate of 5 and 2 were applied, respectively. The capillary temperature was set to 300 °C, the S-lens RF level was set to 68, and the aux gas heater temperature was set to 100 °C.
UV-vis spectra were measured on a Shimadzu UV-1800 spectrophotometer, in a dual-beam setup with a solvent reference. The spectra were recorded in toluene or acetonitrile in a quartz cuvette (1 cm). The absorption wavelengths (λ) were recorded in nm.
Water (20 mL) was added to the reaction mixture and extracted with EtOAc (3 × 200 mL). The organic phases were separated, dried over Na 2 SO 4 , and the solvent was removed in vacuo. The crude product was purified via column chromatography (SiO 2 , Hex/EtOAc 98:2) and recrystallized from acetonitrile to obtain 5 as a clear crystalline solid (635 mg, 81%). 1  Silicone UV Sensors: Silicone sensors were prepared at 0.02 wt% of DAE 5. For each 100 g of silicone resin, 20 mg of 5 was ground for 15 min to a fine powder using a mortar and pestle. It was then dispersed in Vario 15 A (silicone rubber kit from Barnes) and sonicated. Vario 15 B (catalyst) was added to the mixture, stirred rigorously, before being poured immediately into the mold to cure for 24 h. Samples were removed from the mould for testing, and stored under ambient conditions, protected from light.
Prototype LED Box for Resetting UV Sensors: A prototype storage box was developed, that could illuminate the printed sensors with green light to revert them to their colorless state. The housing box was designed using Autodesk Fusion 360 software and 3D printed with an Ultimaker S3 Extended using polylactic acid filament ( Figure S4a, Supporting Information). A simple circuit consisting of a 5 V 1A DC power supply, 4 × 1 W LEDs (520 nm emission), and two switches (Jaycar, Australia) was fit inside the boxes ( Figure S4b, Supporting Information), underneath a Perspex sheet, to provide illumination inside the box.
Printed UV Sensors: 225 mg of the white cubic crystals were ground to a fine powder using a mortar and pestle for 15 min, and then added to 860 g of Formlabs Clear V4 resin (Thinglab, Yarraville, Australia). The mixture was stirred and sonicated for 2 h, protected from light, to ensure complete dispersal of the DAE into the resin (0.03 wt%).
Wearable device designs were produced using Fusion 360 (Autodesk) and Rhinoceros 3D (Robert McNeel & Associates) software. To prepare for 3D printing, designs were exported to .STL file format. These were then imported into PreForm (Formlabs) software, which was used to set print position, generate support layout, and set print parameters. All 3D prints were produced on a Form 2 SLA 3D printer (Formlabs, Somerville, USA) with a 250 mW 405 nm laser, with layer resolution set to 50 µm and automatically generated support layout. The printer was set to open mode, which disabled resin cartridge detection, tank detection, and automatic tank refilling. This allowed for the resin mixture to be manually added to the resin tank without automatically filling from the cartridge, which was required due to the low volume of mixed resin available.
Evaluation of Sensors using UV Light Sources: For controlled UV irradiation experiments, all DAE sensors were exposed to either UV-A or UV-B light using a Rayonet photoreactor (Southern New England Ultraviolet Company, Connecticut, USA). For all experiments, a pair of UV fluorescent battens was used for irradiation of sensors, either LZC-UVA (emission centered at 350 nm) or LZC-UVB (emission centered at 310 nm). Samples were irradiated for fixed periods of time, and at a known distance from the lamps. The intensity of light (W m −2 ) for the experimental configuration for both bulbs was measured using a PM400 optical power meter (Thor Labs, New Jersey, USA), allowing for conversion of irradiation time to irradiation dose (J m −2 ), and intensity adjustments for distance from the lamps.
Evaluation of Sensors using Natural Sunlight: For solution measurements, a 1.70 × 10 −4 m solution of 4 was dissolved in toluene. Three clear sided quartz cuvettes were filled with 2 mL of this solution, and exposed to natural sunlight for 15, 30, and 60 s. UV-vis spectroscopy was measured on each solution after exposure. The sample irradiated for 60 s was irradiated for a further 60 s (120 s) total, and the UV-vis spectrum measured again. This experiment was conducted at QUT Gardens Point Campus, Brisbane, Australia (−27.477247, 153.028106), on the 23rd March 2022, between 12:00 and 1:00 pm AEST, with a measured UV index of 8.9 at this time (ARPANSA, see below).
For exposure of 3D-printed sensors, a collection of different devices was exposed to natural sunlight for 90 s, until no further discernible color change was occurring. Samples were photographed, then exposed www.advmattechnol.de Adv. Mater. Technol. 2023, 8,2201918 to green light for 90 min to reset the color using the prototype LED box (4 × 1 W LED, 520 nm) and photographed again. This experiment was conducted at QUT Gardens Point Campus, Brisbane, Australia (−27.477247, 153.028106), on the 16th August 2022, between 1:00 pm and 2:00 pm AEST, with a measured UV index of 4.1 at this time (ARPANSA, see below).
UV dose for all sunlight experiments was estimated using a simple excel calculation tool. [31] Validation of estimated values was achieved by comparing predicted UV index values from the tool to reported values from the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA, https://www.arpansa.gov.au). The total estimated UV dose from 290 to 400 nm, reported as W m −2 was used to convert time exposed to dose.
Quantification of UV Sensors using Image Analysis: Following irradiation of solutions of DAE, silicone sensors, or 3D-printed sensors, devices were either photographed using the standard camera settings on a Samsung smartphone (Android), or scanned using a flatbed scanner. Care was taken to ensure consistent lighting between images, and no shadows on the samples. For comparison between individual images, control samples with and without irradiation were included in the images as color references.
Quantitative image analysis was performed using either Image-J [32] or Fiji [33] and statistical analysis was performed using Graphpad Prism. RGB images were imported into Image-J or Fiji and cropped to the relevant size. The images were converted to a three color RGB image, and the pixel intensity was inverted so that an increase in the pink coloration correlated with an increase in pixel intensity for the relevant channels. A similar process was used for the HSB color space. Rectangular regions of interest were drawn over the sensor, and the mean pixel intensity in that region measured. Measured values were exported to Graphpad Prism for graphical presentation and statistical analysis. Preliminary method development found the green channel for RGB analysis provided the most direct and reliable quantitation of color change and was used exclusively in subsequent experiments.
For cycling experiments, to ensure consistency between images, a positive reference of an irradiated sample, and a negative reference of an unirradiated sample were included in each photograph. These references were analyzed using the same protocol as all other samples. The mean pixel intensity for cycled sensors was min-max normalized using the following equation

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.