Blue Transparent OLEDs with High Stability and Transmittance for Modulating Sleep Disorders

Optoelectronics devices utilizing organic light‐emitting diodes (OLEDs) are emerging as new platforms for healthcare applications. In particular, wearable optoelectronics such as visual stimulus systems offer a distinctive advantage to intervene in and improve sleep disorders. In this study, two improvements are proposed for transparent OLEDs (TrOLEDs) that will be critical for visual applications. First, zinc sulfide with high surface energy and a high refractive index is explored as a seed and capping layer. An ultra‐thin silver cathode of 8 nm is demonstrated to be feasible in TrOLEDs, and luminous transmittance of 91% is achieved. Second, in general, achieving the operational stability of TrOLEDs with high transmittance is challenging due to the vulnerability of thin electrodes. By introducing a doping process to the electron transport layer, a lifetime comparable to that of control OLEDs with thick cathodes (>90%) is secured. Last, a preclinical model using blue light is proposed to modulate sleep patterns. Melanopsin is stimulated at the highest level of sleep desire, reducing non‐rapid eye movement sleep duration in mice by up to 14%. Based on these results, the proposed TrOLEDs are promising candidates for modulating sleep disorders such as insomnia and narcolepsy–cataplexy with the convenience of wearable form factors.

transparent electrodes has recently advanced to a great degree: in addition to transparent conductive oxide (TCO), [31] many alternative technologies have been explored including graphene, [32] carbon nanotube, [33] conductive polymers, [34] silver nanowire, [35] and dielectric-metal-dielectric (DMD) multilayers. [36][37][38] Among the electrodes for TrOLEDs, in particular, the DMD structures do not employ plasma processes for fabrication, and thus the major organic layers can be protected without utilizing a thin protective layer, [31] and an in situ process, which decreases the fabrication steps, is available. It is also possible to secure high transmittance in the desired wavelength spectrum by controlling the thickness of the dielectric layer having various indices in the evaporation process. As a result, many studies achieved low sheet resistance and high transmittance; however, the DMD electrodes still have challenges in terms of efficiency and stability when they are used in actual TrOLEDs. It is very demanding to ensure high transmittance and device stability simultaneously, and optoelectronic devices with a shortened operational lifetime could limit their use for healthcare applications.
In this study, improving the transmittance and stability of TrOLEDs, which are both important for visual applications, is explored. Notably, unlike previous transparent electrode studies, the researchers sought to achieve both characteristics simultaneously in blue TrOLEDs by controlling the seed layer, capping layer, and carrier behavior. In addition, a preclinical mouse model using blue light is also suggested to modulate sleep patterns by stimulating melanopsin at the highest desire to sleep in nocturnal animals. The changes in the non-rapid eye movement (NREM) duration observed in our model are expected to be used in clinical trials as conditions that could control sleep disorders in daily life.

Design of Transparent OLEDs
TrOLEDs can prevent risks such as those arising in plasma and post-annealing processes through DMD structures [36][37][38] that only involve evaporation processes and where the evaporated materials can be freely changed to achieve the targeted transmittance. In the TrOLEDs with DMD structures, the metal electrode layer and the dielectric capping layer are key features to control the transmittance. As metal candidates, silver is appropriate, with higher transmittance than aluminum and gold when it is sufficiently thin. [37,39] Further, since it is difficult to achieve perfect surface coverage in a thin metal film, [40] this study focuses on a method to induce a 2d continuous film in the silver layer to lower the sheet resistance and increase the carrier injection. The silver layer generally follows the 3d-island formation mechanism named Volmer-Weber growth, as shown in Figure 1a, and therefore it is necessary to apply a seed layer to promote the wetting effect and induce layer-by-layer growth to lower the percolation thickness.
The surface coverage primarily depends on the surface energy of the bottom layer and a proper seed layer should be implemented. In the case of zinc sulfide (ZnS), since it has high surface energy among materials that can be deposited continuously in the evaporator, [37] ZnS was applied as the seed layer and a thin silver film with low sheet resistance without limiting the performance was realized. The results of the sheet resistance when a 1 nm or thicker seed layer of ZnS was added are shown in Figure 1b. A minimum thickness of 15 nm was required to achieve a sheet resistance of less than 10Ω Sq −1 without the seed layer. In contrast, when the seed layer with high surface energy www.advmatinterfaces.de was utilized, the thickness of the silver film had a sheet resistance of 6.4Ω Sq −1 even at a thickness of 8 nm, demonstrating that it was suitable as the electrode for TrOLEDs.
The percolation thickness and the number of grain boundaries changed depending on the existence of the seed layer, which affects the transparency, as shown in Figure 1c,d. The 2D continuous silver film of 8 nm was well-formed when a seed layer was incorporated, thereby having a sheet resistance of less than 10 Ω Sq −1 and exhibiting higher transmittance. On the other hand, the transmittance decreases rapidly even with a very thin silver layer if there is no seed layer. In addition, since the loss of transmittance is greater than the advantage of the sheet resistance obtained by thickening the electrode, this study confirmed that a silver layer of 8 nm was finally suitable when ZnS was used as the seed layer.
In contrast to the resistance of the 2d continuous film determined by Matthiessen's law, non-ideal variables such as surface roughness, bulk resistivity, and grain boundaries mainly contribute to the resistivity of the transparent electrodes. Therefore, grain boundaries and surface roughness were measured, as shown in Figure 2, and the results demonstrated the validity of the 8 nm silver electrode. In the absence of the seed layer, as shown in Figure 2a, the 2d-layer growth model of the thin film was not observed through scanning electron microscope (SEM) images, in contrast with Figure 2b, which had the seed layer of ZnS. It was demonstrated that the grain boundaries of a 3d-island model were found if there was no seed layer, even at the same thickness. Further, since the smoothness of the layer affects the performance of the silver thin film, the surface roughness was measured using an atomic force microscope (AFM) and the root mean square roughness was reduced from 1.153 to 0.592 nm. In conclusion, the silver film with thickness of 8 nm, verified by both surface analysis methods, exhibited sufficiently low sheet resistance to maintain the charge balance in the OLEDs at high transmittance, and it maintained a uniform E-field and surface potential with a roughness of less than 0.8 nm (10% of 8 nm).
In order to improve the transmittance of the DMD structures, it is not only important to apply the thin metal electrode as before, but also necessary to control the reflectance of the metal electrode optically. [37,41] In general, the reflectance of a multilayered film can be controlled by harmonizing the refractive indices, and therefore high transmittance was achieved and the stability of the silver electrode was enhanced by controlling the dielectric capping layer. When light passes through a multilayered thin film such as OLEDs, it produces a transmitted and reflected electromagnetic (EM) field at the boundary of each film layer, as illustrated in Figure 3a. The light is finally emitted by the interference of various fields in layers, that is, the transmittance of multilayered thin films is calculated by considering the boundary conditions and phases of the EM field in every layer. The transmittance simulation estimates the changes in admittance of the multilayered film as each film is deposited and finally calculates the ratio between incident and transmitted EM field. [41] According to the characteristic method, reflection from DMD electrodes in TrOLEDs can be minimized efficiently when an outer dielectric layer with a high refractive index is used as a capping layer. As illustrated in Figure 3b, the transmittance of the 8 nm electrode was simulated with three representative materials as the capping layer. N,N′-Di(1-naphthyl)-N,N′-diphenyl-(1,1′-biphenyl)-4,4′-diamine (NPB), which has a general refractive index as an organic material, molybdenum trioxide (MoO 3 ), which has a higher refractive index, and ZnS, which has one of the highest refractive indices among thermally vaporized materials, were applied and it was confirmed that the reflectance significantly decreased when a high refractive index material was used. Further, the reflectance can be changed repeatedly as the thickness is increased, and the first optimal thickness for maximum transmittance would be in a range of 20-30 nm. Therefore, this study selected ZnS as a core material, and efforts were made to simplify the fabrication and to achieve the highest possible transmittance in TrOLEDs by utilizing it as both the seed layer and the capping layer.

www.advmatinterfaces.de
In order to apply the thin silver and ZnS layers to TrOLEDs, a transmittance simulation of the entire OLED film was performed using the refractive indices and extinction coefficients of each material. As a result, the feasibility of the electrode structure with a thin silver cathode of 8 nm and a ZnS capping layer of 27.5 nm producing the maximum luminous transmittance was confirmed, as shown in Figure 3c and Figure S1, Supporting Information. The measured transmittance spectrum in the visible region of the film was reasonably consistent with the calculated value ( Figure 3d). Although the overall flatness and the transmittance of TrOLEDs at the short wavelength region were slightly insufficient compared to that of the 150 nm-thick ITO film, the luminous transmittance of the total OLED film including the ITO anode was 91%, which is comparable to that of an ITO film.
In conclusion, TrOLEDs with a DMD electrode were designed to achieve both high transmittance and low sheet resistance with an ultra-thin silver layer. As a result, the electrode and the capping layer were well harmonized, and thus transmittance of the entire OLED film equivalent to that of a single ITO film was realized. In particular, by applying ZnS as the core material, our study demonstrated one of the highest luminous transmittances of films among full TrOLEDs reported to date, and the proposed structure can be applied to an actual wearable TrOLEDs platform such as eyeglasses, as shown in Figure 3e and Figure S2, Supporting Information.

Enhancement of the Operational Lifetime in the Transparent OLEDs
In general, it has been challenging to achieve operational stability of TrOLEDs with high transmittance due to the vulnerability of thin electrodes. This study thus aims to minimize electrochemical degradation by controlling excess carriers in the thin metal electrode and achieving improvement in the operational lifetime of the TrOLEDs. [42] As discussed in the previous section, the device structure in this study to improve the The results of the transmittance simulation in full OLED devices by applying NPB, MoO 3 , and ZnS as the capping layer. c) The maximum luminous transmittance according to changes in the ZnS capping layer thickness. The simulation is conducted utilizing thin-film optics in the multilayered film. d) The transmittance spectrum in the visible region by applying the optimal thickness at the maximum transmittance. The ITO transmittance is also illustrated for comparison with full OLED devices. e) Photographs of the blue transparent OLEDs with a luminous transmittance of 91%. Emission from the TrOLED appears white rather than blue; this is because the exposure setting that shows the background together with the TrOLED to illustrate its transparency is relatively high for the OLED itself.  Figure 4b. The reflectance, which was suppressed as much as possible in the vertical direction, increased as the angle changed, and in particular, the reflection in the long wavelength region changed more than in the short wavelength because the luminous transmittance is higher when the spectra match well with the luminosity function, which peaks at the wavelength of 555 nm. As a result, the angular spectrum remained almost constant when the observation angle varied. The detailed performance of the blue TrOLEDs for emission in both top and bottom directions is shown in Figure S3, Supporting Information.
In OLEDs, the flow of excess carriers (electrons or holes) to the organic layer induces defects, and problems in the emission layer (EML) affect the light emission of the device. Further, one kind of carrier should flow to both the electron transport layer (ETL) and hole transport layer; however, electron mobility is insufficient compared to hole carriers in general, and in the case of TrOLEDs, the cathode is also thin, which seriously reduces electron injection. As a result, excessive carrier entering the opposite transport layer because of imbalanced charge in TrOLEDs causes unintentional oxidation or reduction, and the irreversible layers such as Alq 3 could be degraded, [42] as illustrated in Figure 4c. This lowers the stability of the device, and it becomes arduous to drive the transparent device continuously for various healthcare applications. In other words, previous studies achieved high transmittance at the expense of carrier injection, but the carriers and excitons were not confined well, resulting in degradation of both efficiency and the operational lifetime of the TrOLEDs. In this work, Cs 2 CO 3 was utilized to compensate for the decrease of electron injection in the TrOLEDs. [43,44] However, this was insufficient to enhance electron injection in a thin silver cathode compared to a thick electrode. Therefore, this study proposed a new barrier structure between the EML and ETL created by lithium doping to confine excess carriers.
Among materials that can donate lithium ions to the ETL and can be thermally evaporated, there are various candidates such as lithium fluoride (LiF), 8-Hydroxyquinoline lithium (Liq), lithium hydride (LiH), lithium azide (LiN 3 ), and lithium carbonate (Li 2 CO 3 ). In particular, LiH has an advantage in codeposition with other electron transport layers in terms of its decomposition temperature and purity, [45] and this study thus utilized 3 wt% of LiH as a doping material to Alq 3 . The doping process of lithium ions was evaluated by making an electrononly device, and the exact device configuration was ITO/ Alq 3 /Alq 3 :LiH/ZnS/Ag (8 nm). Based on the results shown www.advmatinterfaces.de in Figure 4d, this study demonstrated that electron injection was improved on the doped device compared to the undoped device, and the carriers could be controlled by the band barrier of shifted Alq 3 .
The results of applying LiH doping to ETL (Alq 3 :LiH (3 wt%)) in the proposed TrOLEDs structure are illustrated in Figure 4e. The current density-voltage-luminance (J-V-L) characteristics of the undoped and the doped device were not significantly changed because of the tradeoff effect between the energy barrier and electron density of the LUMO. As a result, although the electron injection did not increase, the energy barrier of the HOMO between the EML and ETL was increased since the energy band of Alq 3 is moved by 0.6 eV, [45] and this would block the excess holes from EML. Figure 4f illustrates the improvement of the operational lifetime in the TrOLEDs before and after doping. For devices without the doping process, ETL degradation occurred due to excessive hole carriers, and the operational lifetime of the device was degraded. Therefore, the time taken to reach 80% of the initial luminance (LT 80 ), which is a representative lifetime for healthcare applications, was less than 150 h (@1000 cd m −2 ). In contrast to undoped devices, when excess holes were prevented in the ETL by doping, the lifetime increased more than twofold to 350 h (LT 80 ) under the same luminance condition (@1000 cd m −2 ). In addition, it turned out that this lifetime corresponds to ≈90% of the control blue OLEDs with a thick aluminum cathode. This result was demonstrated even though TrOLEDs were operated with a twofold higher current level than the control devices to achieve the same level of luminance conditions.
In conclusion, starting from the film formation of a thin silver cathode to the improvement of electrical stability using a doping process, the proposed TrOLEDs were well designed to solve the problems of transparent electrodes. As a result, this study demonstrated one of the longest operational lifetimes among the reported TrOLEDs with high transmittance.

Effect of the Emission Wavelength of OLEDs on NREM Sleep Duration
Previous studies have shown that light-based modulation of hormones such as melatonin via the visual pathway could control the circadian rhythm in mammals as a long-term effect, and mechanistic investigations have led to the discovery of melanopsin. [21][22][23] They have reported that blue light harmonizes with the response wavelength region of melanopsin, which affects the circadian rhythm as an indirect effect. [46] Furthermore, recent studies have proposed new protocols that demonstrate a direct effect between NREM sleep duration and light, [25] which mainly affect insomnia and narcolepsy-cataplexy, as illustrated in Figure 5a. [47][48][49][50][51] However, since the direct effects of light stimulation vary significantly depending on the experimental protocol, there are still challenges related to light sources and sleep stages of stimulation. Therefore, this study proposed a new protocol that provides constant blue light stimulation during Zeitgebertime (ZT) 0 to 2, which shows the highest desire for sleep. The proposed protocol could be expected to shed light on the direct effect (short-term effect) on future human clinical trials because mice are nocturnal animals, and natural light promotes their sleep.
The animal experiment consisted of a mouse, an analysis system to determine sleep stages, and a blue light source system, and detailed surgical procedures for preparing the mouse models and analyzing the data are described in the Experimental Section.
For the light source system, as illustrated in Figure 5b, OLED panels (LL081RR1-54P2, LG Display Co. Ltd, Korea) were attached to four sides of the mouse cage to provide a stable and uniform environment to minimize the variables of the animal experiments. Blue light stimulation was implemented by OLEDs with color filters (Lee Filters) to demonstrate the relationship between melanopsin and sleep duration. In particular, as shown in Figure 5c, although the emission spectrum of the blue TrOLEDs proposed in the previous section and the OLED panels with filters are not exactly the same, the proposed TrOLEDs satisfy the requirements for the experimental irradiance, as described in Table 1 and Figure S4, Supporting Information. In detail, the illuminance conditions of each light source are converted to the amount of irradiance stimulating the melanopsin. Assuming that TrOLEDs in the form of a wearable platform such as glasses operate at about 1000 cd m −2 , an illumination of about 1740 lux could be calculated using a light simulation. [52] In conclusion, although attaching a light source in front of a mouse has not been tested due to the instability of animal experiments, the proposed blue TrOLEDs are expected to be strong candidates for future human studies using a wearable platform since they can stimulate melanopsin more effectively than filtered panels.
The illuminance of white light (daylight) and blue light (stimulation) was designed to be 100 lux at the center of the cage for both in consideration of the field of view in mice. The experiment was managed using a soundproof sleep chamber housing (Sontek, Korea) to maintain the atmosphere of the cage. In addition, a preliminary study was conducted using commercially available LED arrays to verify the new protocol using constant blue light at ZT0 to ZT2, as shown in Figure S5, Supporting Information.
The new experiment protocol, as illustrated in Figure 5d, consists of electrode implantation surgery, a post-surgery stabilization period of at least 1 week, baseline recording for 3-5 days, and stimulation. Mice were housed under artificial white daylight (ZT0-ZT12) and electrophysiological recording was conducted for 10 h every day (ZT0-ZT10). A fast Fourier transform (FFT) was used to determine the baseline sleep duration for NREM and rapid eye movement (REM) sleep. The baseline values for NREM and REM sleep were 55% and 5.5%, respectively. [53] Blue light stimulation was delivered for 2 h (ZT0-ZT2) on a single day after baseline recordings. For the remainder of the day (ZT2-ZT12), regular white daylight was provided. The baseline data of each mouse were used to evaluate the changes in the NREM sleep duration in the stimulation period. Further, all changes in the sleep duration were the results of the day of stimulation, and although the observation was continued during days after the first stimulation, meaningful results were difficult to confirm since the baseline was different for each individual after stimulation.

www.advmatinterfaces.de
The main study using OLEDs demonstrated a 14.1% decrease (two-tailed unpaired t-test, ***p < 0.001, N = 8) in NREM sleep duration compared to the baseline, as shown in Figure 6. The standard deviation was 6.53 and the average NREM sleep duration on the day of blue light stimulation was 85.4%. Thus, these results show the effect of constant blue light on NREM sleep duration in the proposed mice when their desire for sleep is high.
In conclusion, this study demonstrated a significant decrease in NREM sleep duration through the proposed protocol (ZT0 to ZT2, 2 h stimulation, and non-flickering blue light) in vivo experiments. Further, we expect that the proposed OLED eyeglasses with high transmittance described in the previous section will provide greater clinical compatibility and degree of freedom due to their design flexibility and transparency for integration into healthcare devices.

Conclusion
In summary, in this study, we proposed blue TrOLEDs with high stability and transmittance, and introduced a new protocol for sleep management applications. For the blue TrOLEDs, we adopted an optimal, thermally-evaporated DMD as their top cathode, which allowed us to avoid potential plasmainduced damage to organic layers that can be expected in the Figure 5. a) Schematic diagram of the biological mechanism related to circadian rhythm by blue light, and its indirect (long-term circadian) and direct (short-term sleep duration) effects. b) Schematic diagram of the preclinical animal model utilizing OLED light sources and photographs of the inside of the experimental cage. c) The relative sensitivity spectrum of melanopsin and the EL spectra of the OLED panel with a blue color filter and the blue transparent OLEDs under study. d) The experimental protocol with blue OLED stimulation: daily routine from Zeitgeber time 0 to 24, and week schedule from surgery to stimulation. Calculated under mouse conditions considering pupil size by pupillary light reflex, size of TrOLED = 10 mm × 10 mm, pupil-to-OLED distance = 5 mm, and pupil area = 0.16 mm 2 .

www.advmatinterfaces.de
case of RF-sputtered TCOs. Silver and ZnS were suggested as a metal layer and the optical capping layer for the DMD electrode, respectively, to achieve transmittance that is as high as possible. In addition, a thin ZnS layer was utilized as a seed layer to improve the wetting of silver to form a thin yet smooth and dense film even at a relatively low thickness. Based on the optical design of the multilayered structure, the luminous transmittance of the proposed blue TrOLEDs was demonstrated to be as high as 91%, which is similar to that of a single ITO film of 150 nm. In addition, we finally demonstrated OLED eyeglasses with high transmittance which we believe can unlock the full potential of our proposed TrOLEDs for visual stimulation applications.
For wearable optoelectronic devices to be used in people's daily lives, it is crucial to improve their operational stability as well as transmittance. However, the LT 80 operational lifetime of the proposed TrOLEDs was less than 200 h whereas the control blue OLEDs with a thick aluminum electrode exhibited an LT 80 operational lifetime of about 400 h. Compared to the control device, the degradation of the TrOLEDs was attributed to Alq 3 , used as an ETL. Specifically, its limited performance reduced electron injection and the concomitant charge imbalance accelerated the reduction of the operational lifetime. A method was thus devised for Li + doping in the ETL, and as a result, the operational lifetime at LT80 was enhanced more than twofold to 350 h, which approached the operational lifetime of the control device (>90%) with thick metal electrodes.
In addition, we investigated the relationship between constant blue-light illumination and NREM sleep duration. The discovery of the reaction of melanopsin to blue light enabled the research on the control of sleep duration. We proposed a new protocol to investigate the direct effect on melanopsin by giving constant blue light stimulation at the stage of the highest desire for sleep in the nocturnal mouse. Although OLED panels with color filters were utilized to minimize the side effects in animal experiments, as shown in Table 1 and Figure S4, Supporting Information, we confirmed that the proposed blue TrOLEDs can also provide stable and sufficient amounts of stimulation to melanopsin. The results of the present study indicate that illumination of blue light can have an immediate effect of suppressing sleep duration by over 10%.
In conclusion, this study proposed blue transparent OLEDs as eyeglasses form with more stable and higher transmittance for wearable sleep management devices. The relationship between NREM sleep duration and constant blue light illumination covering the absorption spectrum of melanopsin was also investigated through a new protocol for the nocturnal mouse. From the results of this study, it is expected that the proposed blue TrOLEDs will become a viable candidate to modulate sleep disorders such as insomnia and narcolepsy-cataplexy with the convenience of wearable form factors.

Experimental Section
Animal Care and Use Statement: It was declared that all animals used in this study were reviewed and approved by the IACUC (KA2019-25) at the Korea Advanced Institute of Science and Technology (KAIST). In addition, H. Chae, who led this animal experiment, Y. Jo, who helped conduct the experiment, and corresponding author K. C. Choi had completed the mandatory animal care and use certification training. A total of 12 male C57BL/6NTac wild-type mice (8-10 weeks old) were used for in vivo experiments. The mice were purchased from DBL Co., LTD and they were certified (ISO9001 / ISO14001) and passed AAALAC standards. All animals were housed in separate cages in a soundproof sleep chamber (Sontek, Korea) with lights on at 10:00 and off at 22:00. The illuminance of light during the adaptation period was maintained at 100 lux to establish the same environment as that of the experiment in this study.
Mouse Surgery Procedure: The exact procedure of the surgical method is shown in Figure S6, Supporting Information. First, brief inhalation anesthesia was performed with isoflurane, and for longterm anesthesia during surgery, mice were injected intraperitoneally with Avertin (20 µL g −1 ) and observed until fully sedated. A small amount of analgesic was then injected for the operation and the mice were head-fixed on a stereotaxic frame (RWD Life Science Co., LTD., China) and the fur on the top of the head was removed using a pair of stainless-steel medical scissors. After removing the fixed mouse's scalp with medical scissors, a povidone-iodine solution was used to remove the slippery scalp membrane. A total of three electrodes were then inserted: 1) EEG signal, 2) EMG signal, and 3) common electrical reference. The EEG and reference electrodes were inserted into the skull at ML:2, AP:−2 mm; and ML:−2, AP:−5 mm from the bregma, respectively. The EMG electrode was inserted using a 26-gauge needle and a very thin wire across the trapezius at the back of the neck and tied so that the movement of the trapezius can be received as an electrical signal. Finally, all three electrodes were covered with dental cement and fixed to the skull to minimize infection and reduce damage to the integrated probe due to mouse movement. After surgery, there was a recovery period of at least a week. If there was an abnormal appearance, such as an abnormal posture or very rough hair, or if there was a weight loss of 20% or more compared to the normal weight without ingestion of feed or water, euthanasia was induced using isoflurane according to ethics protocols.
Analysis of EEG and EMG Data: The measured raw data were analyzed, as shown in Figure S7, Supporting Information. The EEG/EMG signals were sampled at a rate of 1 kHz and digitally amplified with a gain of 1000 using a bio-potential acquisition device (RHD2000, Intan Technologies, CA, USA). The signal was then filtered with a 0.1 Hz low-pass filter, www.advmatinterfaces.de 7.5 kHz high-pass filter, and 60 Hz notch filter. The electrophysiological signals obtained through the bio-potential acquisition device are shown in Figure S7a, Supporting Information. During sleep, only the heartbeat signal was detected through the EMG signal and there was no noise in the EEG signal. Conversely, when mice were awake, they generated a large signal in the EMG data. The electrophysiological signals were analyzed through a sequence diagram, as illustrated in Figure S7b,c, Supporting Information. A custom-written MATLAB program was used to partition the signal into 6 s epochs and a FFT analysis was run for each incoming epoch of the real-time signal. For every epoch, the FFT analysis determined the sleep/wake state of the mice by comparing the EEG/EMG power spectrum of the signal in the frequency domain. EEG delta waves (0.5-4 Hz) are dominant during NREM sleep, theta waves (4)(5)(6)(7)(8) are dominant during REM sleep, and alpha waves (8)(9)(10)(11)(12) are dominant during WAKE states. In particular, in the case of mice, unlike humans, sleep and wake states were repeated rather than continuous sleep. Therefore, although WAKE, NREM, and REM states could be included in each epoch at the same time, the dominant single state was represented in this analysis. In addition, the data used in this study were the result of measuring the behavior of the mouse at ZT0 to ZT10, and it was concluded that the baseline was established when the pattern appeared constantly during the measurement, which was divided into five sections of 2 h.
OLEDs Fabrication: In this study, the OLEDs were fabricated through thermal evaporation in a single vacuum process, and the deposition vacuum state was less than 3.0 × 10 −6 torr. The typical organic layers were deposited at 1 Å s −1 , and the silver layer was also deposited at 1 Å s −1 to control the thickness precisely. The measurement of the operational lifetime was carried out at 20 °C and 60% humidity.

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