Solid‐State and Flexible Black Electrochromic Devices Enabled by Ni‐Cu Salts Based Organohydrogel Electrolytes

Solid‐state black electrochromic devices (ECDs) are promising for smart window applications, particularly when privacy protection and low leakage are required. Herein, a Ni–Cu salts/poly(vinyl alcohol) based organohydrogel electrolyte is developed with superior visible‐light transparency (83.8%), ionic conductivity (0.11 mS cm−1), and mechanical properties (tensile strength: 11.1 kPa, breaking strain: 242.6%). Due to the high viscosity of the organohydrogel electrolyte, a homogeneous Ni–Cu alloy film with a surface roughness of around 11.2 nm can be electrodeposited under −3 V for 5 min, and the resulting black color can be retained for over 350 min with a transmittance increase of only 5% at the voltage‐off state. The solid‐state rigid ECD exhibits an outstanding optical contrast between the transparent and colored states (visible light transmittance: 70.8% vs 0.085%), excellent cycling stability with over 90% retention of optical contrast after 2000 cycles. Finally, a flexible ECD is fabricated with the organohydrogel electrolyte and annealed indium tin oxide (ITO)‐coated polyethylene naphthalate (PEN) films as flexible and durable electrodes. It exhibits good mechanical flexibility with transmittance modulation degradation of 10% after 800 bending cycles and switching stability for 400 cycles with up to 43% optical contrast.


Introduction
An electrochromic device (ECD) controls optical characteristics such as transmittance, reflectance, or emittance in a reversible manner upon electrical bias. This distinctive characteristic enables ECD to be utilized in a variety of applications, including green buildings [1] and electrochromic displays. [2] Until now, the majority of ECDs published in the literature have been focused on the liquid or gel polymer electrolytes, imposing relevant shortcomings such as leakage due to insufficient sealing and poor chemical stability. [3] In order to conquer these essential limitations, many www.advmatinterfaces.de created a novel ECD based on reversible silver electrodeposition employing a SPE comprised of poly(ethylene oxide), perchloride acid, and potassium iodide, which attained a high initial transmittance of 90%. Mello et al. [10] developed solid electrolytes by dissolving copper salts into an animal protein-based gel, and the assembled ECD based on reversible electrodeposition of copper demonstrated a visible-light transmittance modulation of 63.6% and good thermal stability at 200 °C. Nonetheless, the cyclability of the ECDs in the above two reports was restricted to 50 cycles with respective 75% and 45% optical contrast, demonstrating limited service life. Although solid electrolytes have garnered considerable attention due to their outstanding electrochemical stability and low leakage, [11] the limited ionic conductivity and a restricted electrochemical window remain ongoing challenges for solid-state ECDs. Therefore, developing a stable solid electrolyte with a wide potential window and high ionic conductivity is desired for solid-state REDM-based ECDs.
Flexible REDM-based ECD with the solid-electrolyte and flexible electrode is another important direction in meeting a variety of growing demands. [12] Besides the solid electrolytes, developing flexible transparent electrodes is also challenging. ITO is always a viable option for fabricating transparent conductive electrodes of ECDs due to its excellent transparency and conductivity. Nonetheless, most commercially available ITOcoated flexible electrodes, such as ITO-coated poly(ethylene naphthalate) (PEN), are unsuitable for REDM-based ECDs since the ITO nanoparticles may be dissolved into the electrolyte upon application of electric field, resulting in a loss of conductivity. Additionally, due to the non-uniformity of the magnetic field during the sputtering process, the as-deposited ITO thin films are typically non-uniform and may contain structural flaws. [13] Therefore, improving the stability and conductivity of flexible ITO/PEN electrodes is also crucial for developing high-performance flexible REDM-based ECDs.
According to our previous work, an ECD based on reversible Ni-Cu electrodeposition is able to fulfill good cycling stability and large optical contrast. [14] Herein, an optimal solid-state Ni-Cu salt/PVA-based organohydrogel electrolyte was prepared with a mixture of DMSO/water (80/20 in weight) as the solvent via a freezing-thawing method. A solid-state REDM-based black ECD containing this organohydrogel electrolyte was assembled to fulfill outstanding cycling stability for 2000 cycles with up to 53.1% transmittance modulation at 550 nm. Furthermore, the device can reach a great visible-light transmittance modulation of over 70% in 2 min, achieving superior color neutrality (C* = 1.33). Additionally, a flexible REDM-based black ECD with the above organohydrogel electrolyte and upgraded flexible electrodes (annealed ITO/PEN films) was developed to achieve switching stability for 400 cycles and good mechanical flexibility with optical contrast degradation of only 10% after 800 bending cycles at the curvature radius of 20 mm.

Optimization of Organohydrogel Electrolytes
PVA is chosen as the polymer matrix of the solid electrolyte in this work due to its non-toxicity, high ionic conductivity as well as excellent mechanical integrity. [4,15] There are two general methods to induce the gelation of PVA: chemical and physical cross-linking techniques. [15b] Chemical cross-linking typically requires the inclusion of a cross-linker (e.g., glutaraldehyde [16] ) or irradiation (e.g., gamma ray [17] ), which may result in undesirable side reactions with other additives (e.g., metal salts) and the introduction of hazardous components. Physical cross-linking utilizes the intermolecular bonding of PVA chains and results in pure hydrogels. One successful example is the freezing-thawing (FT) of aqueous PVA solution based on the mechanism of phase separation and crystallites formation, [18] and the resulting hydrogel can be customized in terms of transparency, degree of crystallinity and mechanical properties, by controlling the concentration of PVA, FT cycle numbers, and so forth. [15b] Besides water, DMSO was also used as the solvent for PVA-based gel electrolytes due to its good stability, wide electrochemical window as well as low volatility. [14] Nonetheless, the PVA gel electrolyte with only DMSO as the solvent suffers from long gelation time (or gelation cannot be observed), low ionic conductivity, and poor mechanical properties, which potentially affect the transmittance modulation and stability during cyclic testing when it works as the electrolyte for ECDs. [19] Thus, it has been hypothesized and proved that the addition of water in the electrolyte can accelerate the gelation process and enhance the conductivity as well as mechanical properties of the gel due to the enhanced degree of crystallization. [19b,20] In this work, a modified FT method as illustrated in Figure 1a is used to fabricate Ni-Cu salts-based organohydrogel electrolytes for solid-state REDM-based ECDs. Initially, 8 wt% PVA powders were dissolved into a liquid electrolyte of NiCl 2 /CuCl 2 / LiClO 4 in mixed water and DMSO solution (See Figure S1, Supporting Information). The fraction of DMSO in the solvent was varied from 100 to 60 wt.% in order to investigate the role and optimal concentration of water. Then the mixtures were cooled in a mold, debubbled under vacuum, frozen at −20 °C for 20 h, and finally thawed at room temperature, resulting in solid-state organogel/organohydrogel electrolytes with different transparency. These gel electrolytes are denoted as FT-DMSO x , where x is the weight percentage of DMSO in the mixture of DMSO and water.
Transparency is a key parameter for ECDs to ensure exceptional optical modulation. As demonstrated in Figure 1b, the visible-light transmittance of both FT-DMSO 80 and FT-DMSO 90 organohydrogels (84.3% and 83.8%, respectively) is comparable to that of FT-DMSO 100 (79.1%), while FT-DMSO 60 organohydrogel shows the least transparency of only 39.4% probably due to the high crystallinity of PVA. The ionic conductivity of the gel electrolytes is also important for solid-state ECDs and can be measured from the electrochemical impedance spectra (EIS) (Figure 1c). The calculated ionic conductivity of organohydrogel electrolytes increases from approximately 0.04 to 0.14 mS cm −1 as DMSO fraction decreases from 100 to 60 wt.% due to the exceptional capability of water in dissolving metal salts and dissociating metal cations from their counter anions. [21] Additionally, a wide electrochemical window of the solvent in the electrolyte is essential to ensure the stability of ECDs by preventing undesired side reactions. To investigate the electrochemical windows of the mixed solvents, the equivalent amount (1.5 mmol) of inert metal salt (NaCl) was employed www.advmatinterfaces.de in place of NiCl 2 and CuCl 2 as a reference, which avoids the possible redox reactions of Ni 2+ or Cu 2+ during linear sweep voltammetry process that affects the actual current density contributed from the solvents. A cut-off current density of ±0.1 mA cm −2 is employed to determine the electrochemical window, as shown in Figure S2 (Supporting Information). [22] Due to the much broader electrochemical window of DMSO (4.4 V [23] ) than water (1.23 V [24] ), the higher DMSO fraction in the mixed solvents and wider the electrochemical window, better stability has been achieved during the reversible electrodeposition and dissolution process.
The mechanical performances of the organohydrogel electrolytes were determined by tensile testing. The typical tensile curves are shown in Figure S3 (Supporting Information). Figure 1d reveals good tensile strength and breaking strain of the organohydrogel electrolytes, especially for the FT-DMSO 60 one (25.2 kPa and 320% respectively). The tensile strength increases as the DMSO fraction decreases from 90 wt.% to 70 wt.%, reaching ≈5.2, 11.1, and 18.2 kPa, respectively. The tensile measurement of FT-DMSO 100 organogel was not successful due to its inability to form a free-standing film. Figure S4 (Supporting Information) shows the real-time photos of the FT-DMSO 80 organohydrogel under the initial, 100% as well as 240% tensile strains. The dynamic mechanical analysis of organohydrogels is essential to identify whether they are solid-state electrolytes, shown in Figure 1e,f and Figure S5 (Supporting Information). The storage modulus (G') and loss modulus (G'') of an elastic material is defined as the capacity to store energy and the amount of energy loss, respectively, and the ratio of the two (G''/G') refers to loss tangent (tan δ). [25] Solid-like gels need to meet two criteria: the storage modulus has a distinct plateau at the low-frequency region, and the loss modulus is less than the storage modulus. [26] As seen in Figure 1e, the storage modulus of organohydrogels shows a noticeable plateau when the angular frequency is around 0.1 to 1 rad s −1 , while that of FT-DMSO 100 organogel exhibits a minor increase with frequency. Additionally, the loss tangent values indicated in Figure 1f are all smaller than 1, demonstrating that the storage modulus is always greater than the loss modulus. Therefore, the organohydrogel electrolytes with DMSO fraction ranging from 60 wt% to 90 wt% can be regarded as solid-like electrolytes. To demonstrate the role of FT process, the storage and loss moduli of DMSO 80 electrolyte www.advmatinterfaces.de prior to and after FT treatment are compared in Figure S6 (Supporting Information). The storage modulus of DMSO 80 electrolyte without FT process is always smaller than loss modulus, whereas the storage modulus of FT-DMSO 80 organohydrogel not only has larger values than loss modulus but also exhibits a plateau over 0.1 to 10 rad s −1 , demonstrating the successful transformation from a liquid-like electrolyte to a solid-like one after FT cross-linking process. Viscosity of the electrolyte is a crucial factor in determining the sealability of ECDs. As illustrated in Figure 1g, FT-DMSO 60/70/80 organohydrogel electrolytes show significantly higher viscosities (>0.26 McP at shear rates < 1 s −1 ) than FT-DMSO 100 organogel (<50 kcP), implying much better leak-proof capability.
A PVA gel mainly consists of an amorphous phase formed by randomly distributed PVA chain segments in the solvents and a crystalline phase formed by regularly and tightly stacked chain segments. [27] To investigate the mechanism of liquid-tosolid phase transition induced by the FT process, the crystalline phase in the gel electrolytes was characterized by X-ray diffraction (XRD) and differential scanning calorimetry (DSC). The XRD profiles of the gel electrolytes are shown in Figure 2a.
It can be seen that the intensity of characteristic peaks of PVA (2θ = ≈22.1) increases obviously as the fraction of DMSO decreases, indicating water-enhanced crystallization of PVA chains during the FT process. Additionally, DSC was employed to evaluate the degree of crystallinity (T m of PVA is around 220 °C). As shown in Figure 2b,c, FT-DMSO 60 organohydrogel achieves the maximum degree of crystallinity of 38.9%, which gradually declines to 22.1% as the DMSO fraction increases. This conclusion is consistent with the results of tensile strength and XRD. These results both indicate that the addition of water greatly promotes the crystallization of PVA chains during the FT process, probably due to that water molecules change the solvation behavior of PVA in DMSO. The evolution of the conformation of PVA chains during the fabrication process can be illustrated in Figure 2e. PVA chains are first dispersed randomly throughout the electrolyte solution. Then, phase separation is induced during the freezing stage as the frozen substances (i.e., water/DMSO ice or salt crystals) push PVA chains closer together, enabling the formation of inter-and intra-chain hydrogen bonds and tiny crystallites as physical cross-linkers. Following thawing, the frozen substances are melted into liquid www.advmatinterfaces.de surrounded by the crystalline PVA network, forming a solidstate organohydrogel electrolyte. The water-enhanced crystallization and mechanical properties of organohydrogels can be explained in Figure 2d. Due to the strong interaction between water and DMSO, the addition of water can break the weak hydrogen bonds between DMSO and PVA and facilitates the formation of hydrogen bonds between the hydroxyl functional groups on the PVA polymer chains during the freezing treatment, leading to small crystallites as the cross-linkers of the solid-state electrolyte.

Transparency and Electrodeposited Metal Films of REDM-based ECDs
The effect of different electrolytes on the transparency and deposition of Ni-Cu alloy films in the REDM-based ECDs is shown in Figure 3a depicts the configuration of a rigid ECD, which consists of two FTO glass electrodes and an organohydrogel electrolyte sandwiched between them. When a highly negative potential is applied, the Cu and Ni ions in the electrolyte receive electrons and become reduced, leading to the formation of black CuNi alloy film on the working electrode. Conversely, the CuNi alloy gets oxidized and dissolves back into the electrolyte under a slightly positive potential, resulting in a transparent-state device. On the other hand, the charge transfer at the counter electrode is balanced by the reversible Cl − /ClO − redox reaction, according to our previous work. [14] The transparency of rigid ECDs containing different organohydrogel electrolytes was determined. As indicated in Figure 3b, the ECD containing FT-DMSO 80 organohydrogel electrolyte possesses the highest visible-light transparency (70.8%). In contrast, the ECD involving FT-DMSO 60 organohydrogel electrolyte has the lowest transparency (27.9%). These results agree well with the transparency of the organohydrogel electrolytes (Figure 1b). The morphology of electrodeposited Ni-Cu alloy films when −3 V is applied for 5 min is shown in Figure 3d and Figure S7 (Supporting Information). The average size of Ni-Cu nanoparticles increases with the rise of DMSO fraction of the gel electrolyte (Figure 3c), with the smallest (≈174 nm) and largest (≈247 nm) values achieved in the FT-DMSO 60 and FT-DMSO 100 based ECDs, respectively. This is because the high viscosity of FT-DMSO 60 organohydrogel (see Figure 1g) slows down the electrodeposition rate and realizes a more compact and homogenous film morphology. In comparison, FT-DMSO 100 organohydrogel exhibits the lowest viscosity (see Figure 1g), which has much less limitation of the growth of nanoparticles, hence resulting in a larger nanoparticle size. The performance comparison of different electrolytes for the REDM-based ECDs is summarized in Table S1 (Supporting Information). Considering the excellent transparency and moderate electrochemical window, ionic conductivity, mechanical properties, and achieved nanoparticle size, the FT-DMSO 80 organohydrogel is the optimal option for the following solid-state REDM-based ECD study. The morphology of electrodeposited alloy thin film from FT-DMSO 80 organohydrogel electrolyte was further observed by atomic force microscope (AFM). Due to the high viscosity of the FT-DMSO 80 electrolyte (over 100 kcP at shear rates < 1 s −1 ), the limited electrodeposition rate inhibits the aggregation of Ni-Cu nanoparticles, leading to a uniform alloy film with a low  (Figure 3f) and thickness of ≈600 nm (Figure 3e). The deposited thin film can efficiently block visible light and deliver a black color of the ECD.

Performance of Solid-State Rigid REDM-Based ECDs
A solid-state rigid REDM-based ECD containing FT-DMSO 80 electrolyte and FTO glasses was assembled for investigating the electrochromic performance. In situ visible-light spectra of the ECD at the original transparent state and colored states under various potentials (−2, −2.5, −3 V) for different durations (60, 120 s) are presented in Figure 4a. At the original transparent state (top inset), the ECD has a high visible-light transmittance of 70.8%, with air as the reference. When Ni-Cu alloy film was electrodeposited on the FTO electrode under −3 V for 120 s (bottom inset), the visible-light transmittance is reduced to less than 0.085%, resulting in an outstanding transmittance modulation of over 70%. Additionally, the solid-state black ECD with the FT-DMSO 80 electrolyte exhibits superior memory retention. As illustrated in the Figure 4b, after applying the same pulse of −3 V for 5 min, the black ECD containing traditional organogel electrolyte with 2 wt.% PVA returns to its initial transparent state in ≈85 min when voltage is off, [14] whereas the

www.advmatinterfaces.de
transmittance of solid-state black ECD increases by only 5% in 350 min and 10% in 523 min, indicating the excellent capability to maintain the colored state at voltage-off state (i.e., memory retention performance). A comprehensive colorimetry investigation of the black ECD was undertaken using CIE 1976 L*a*b* color model, where L* signifies lightness value and varies from 0 (black) to 100 (white), a* and b* signify respective green-red and blue-yellow hues. Because the human eye has difficulty in differentiating colors, the chroma (C*) value calculated from the formula (C* = 2 2 * + * a b ) can be employed to evaluate color neutrality. When C* value is smaller than 10, the objects are regarded as grey (neutral). The solid-state black-colored ECD obtained at the bias of −3 V for 120 s (L* = 0.8, a* = 0.7, b* = 1.13) exhibits a superior color neutrality (C* = 1.33), which is comparable to the state-of-the-art dynamic windows based on REDM (denoted as square shapes in Figure 4c, C* ≈ 3 at colored state). [8b,28] Additionally, compared with the black state of the ECD fabricated with polymer [29] (denoted as circle shapes in Figure 4c) or WO 3 [30] (denoted as triangle shape in Figure 4c) coated working electrode, which always displays a large L* or C* value as shown in Table S2 (Supporting Information), the black colored ECD based on the mechanism of reversible metal electrodeposition exhibits a more visually impeccable black hue and greater color neutrality. This solid-state black ECD with exceptionally low transmittance, superior memory retention and flawless black hue is an excellent option for smart glass applications, particularly where privacy protection and leakproof is required. Figure 4d illustrates the switching stability of the REDMbased black ECD (λ = 550 nm) by applying the potential of −2 V (15 s) followed by 0.5 V (60 s) for 2500 cycles. When the metal nanoparticles are electrodeposited onto the FTO working electrode, a uniform black alloy layer is formed and hinders the light transmission, resulting in a low transmittance. On the contrary, when a reversed potential is applied, the metal alloy film dissolves into the electrolyte, leading to increased transparency. In the first few hundreds of cycles, the device is gradually activated to initiate more uniform deposition thus decreasing the transmittance at the colored state. Following that, there are some remnant metal oxide nanoparticles left on the cycled electrode (as shown in Figure S8, Supporting Information), which will affect the nucleation and formation of homogeneous Cu-Ni alloy layers in subsequent cycles. This leads to a gradual increment in the transmittance at the colored state for prolonged cycling. The response time at various cycles is calculated as shown in Figure 4e. The coloration time decreases gradually during the activation process and reaches a maximum coloration rate of less than 10 s. On the contrary, the bleaching time is prolonged with the increase of optical contrast and the minimum bleaching time is approximately 12 s. After 1500 cycles, the bleaching speed accelerates due to the decrease of optical contrast. Additionally, the coulombic efficiency during cycling testing is always greater than 90% and can reach up to around 97%, indicating good reversibility (see Figure S9, Supporting Information).
A large-scale REDM based ECD (225 cm 2 ) was assembled and demonstrated in its colored state as Figure 4f. To investigate the uniformity of metal deposition in the large-area ECD, the transmittance modulation was measured at the side and center points of the ECD using a bias of −3 V for 5 min. Prior to measurements in Figure 4g,h, 15 cycles of reversible electrodeposition (−3 V/30 s) and dissolution (0.5 V/60 s) have been carried out to activate the newly fabricated device via initiating nucleation on the bare FTO electrode. [8a] It is observed that the transmittance at the side point decreases more rapidly than that at the center point over the first 60 s because the voltage is applied at the side locations. Nonetheless, as seen in the Figure 4g, subsequent electrodeposition brought the transmittance modulation at the center close to that at the corner, resulting in a uniform visual response over the whole window. Additionally, it is found that the modulation of charge density required as a function of time for ECD's side and center points are almost identical, and the curves are almost overlapped as shown in Figure 4h. As shown by the comparison Table S3 (Supporting Information), the solid-state ECD containing organohydrogel electrolyte outperforms the counterpart with traditional organogel electrolyte [14] (our previous work) in terms of surface roughness as well as electrochromic performance, with the exception of a longer switching time. Table S4 (Supporting Information) summarizes the performance of various solid-state ECDs in the literature. Solid-state ECDs based on REDM mechanism are scarcely reported and the devices in those reports exhibited poor cycling stability with just 50 cycles. The majority of solid-state ECDs utilized electrochromic materials (mostly WO 3 ) that rely on the insertion of ions such as H + and Li + . They are typically characterized by excellent initial transparency and rapid switching speeds. Our work demonstrates comparable optical contrast and cycling stability, while eliminating the precoating of electrochromic materials.

Performance of Flexible REDM-based ECDs
Flexible electrodes with excellent transparency, conductivity, and stability are essential for high-performance flexible ECDs. Numerous publications indicate that high-temperature annealing could promote the crystallization of ITO thin film, resulting in more stable ITO thin films on the substrates. [31] Additionally, flexible PEN (T m = 270 °C) can withstand annealing at higher temperatures than flexible polyethylene terephthalate (PET) (T m = 255 °C). [32] Thereby, in this work, ITO/PEN sheets were chosen as the flexible electrodes after being annealed at various temperatures (150 and 200 °C) for 30 min. The XRD patterns of the ITO/PEN sheets are shown in Figure 5a. It can be observed that the pristine ITO/PEN exhibits an amorphous background devoid of visible peaks. As the annealing temperature is increased, the crystalline peaks such as (222) at 2θ = 30.5 °C, (400) at 2θ = 35.4 °C, (440) at 2θ = 51.0 °C, (622) at 2θ = 60.6 °C becomes increasingly visible and intense, suggesting the crystallization of ITO thin film. [33] Due to the fact that the flexible PEN substrate suffers from warpage at 200 °C, the annealing temperature for the following study is fixed at 150 °C. Additionally, as shown in Figure 5g, the surface resistance of ITO/PEN film is reduced after annealing due to the formation of larger crystallites (as shown in Figure 5b,c), which results in less grain boundaries and stacking defects sites, leading to increased carrier mobility. [31d,34] Figure 5d,e depicts the photos of a flexible REDM-based ECD consisting of two www.advmatinterfaces.de annealed ITO/PEN sheets sandwiching the FT-DMSO 80 electrolyte at its transparent and colored states. When a negative potential is provided, a Ni-Cu alloy layer is electrodeposited on the working electrode to give a black color; while the bleached state is achieved when a reverse bias is applied.
To examine the stability of flexible REDM-based ECDs, the transmittance modulation was investigated by applying the same bias of −3 V for 15 s and 0.5 V for 60 s, as exhibited in Figure 5f and Figure S10 (Supporting Information). The flexible ECD fabricated with pristine ITO/PEN electrodes exhibits low optical contrast (<10%) over 400 cycles, and the sheet resistance of the isolated working electrode after cycling could reach over 1 MΩ sq −1 (Figure 5g). In comparison, the ECD constructed with annealed ITO/PEN electrodes demonstrates excellent cycling stability for 400 cycles with up to 43% optical contrast. And the annealed ITO/PEN working electrode retains a superior electrical conductivity with a sheet resistance of around 100 Ω sq −1 after cycling (Figure 5g), agreeing well with the better cyclic transmittance modulation. In order to investigate the performance degradation of flexible ECDs after cyclic testing, the morphology of bleached working electrodes after 400 switches were monitored by SEM. Figure S11a (Supporting Information) depicts some micro-cracks on the cycled pristine ITO/PEN, which are potentially caused by the dissolution of ITO nanoparticles into the electrolyte, leading to the poor conductivity of the electrode (Figure 5g). In contrast, the surface of annealed ITO/PEN working electrode is much more intact after 400 cycles ( Figure S11b, Supporting Information). Additionally, www.advmatinterfaces.de some metal nanoparticles were observed on the cycled annealed ITO/PEN working electrode, decreasing the transparency of the bleached device and affecting the nucleation and formation of homogeneous Cu-Ni alloy layers in subsequent cycles, and consequently leading to narrower transmittance modulation for prolonged cycling. The energy-dispersive X-ray spectroscopy (EDX) analysis ( Figure S11c,d, Supporting Information) shows that ITO (In 2 O 3 :SnO 2 = 9:1) is difficult to be detected from the cycled pristine ITO/PEN, whereas 16.3 atomic% of In element can be detected from the cycled annealed ITO/PEN, indicating the excellent stability and durability of ITO thin film on the PEN substrate after annealing. To conclude, the degradation of the pristine ITO/PEN sheet is due to the gradual damage of the ITO layer, and the post-annealing enables the ITO/PEN sheet obviously to be more stable and reliable for flexible ECDs.
To examine the mechanical stability and flexibility of the flexible REDM-based ECDs with annealed ITO/PEN electrodes, the transmittance modulation (λ = 550 nm) at the original straight state and bending states with various curvature radius (15 mm, 20 mm, 25 mm) is demonstrated in Figure 5h by applying a bias of −3 V for 15 s, followed by 0.5 V for 60 s. It is observed that the four curves are almost identical and overlapped, indicating the excellent functionality of the flexible ECD at various bending states. Additionally, the transmittance of the flexible ECD in both transparent and colored states was measured as a function of bending cycles with a curvature radius of 20 mm. As observed in Figure S12 (Supporting Information), the flexible ECD (4 × 2 cm 2 ) was monitored under the bias of −3 V (coloring process) followed by 0.5 V (bleaching process) after every 100 bends. The transmittance modulation between transparent and colored states dropped by only ≈10% after 800 bending cycles (Figure 5i). It is found that the performance degradation of the flexible ECD after repeated bending is due to the fractures on the electrodes, as shown in Figure S13 (Supporting Information), resulting in the formation of inhomogeneous Cu-Ni alloy layers on the working electrode.

Conclusion
A solid-state REDM-based ECD was successfully assembled with a solid-like Ni-Cu salts/PVA-based organohydrogel electrolyte fabricated by a facile and low-cost freezing-thawing method. The adoption of mixed solvents with an optimal DMSO/water ratio of 4/1 accelerates the gelation process and enhances the ionic conductivity and mechanical properties of the organohydrogel electrolyte, while maintaining high transparency. The solid-state black ECD can undergo a continuous switching between bleached and black states with a transmittance modulation of up to 53.1% for more than 2000 cycles. The black ECD achieves outstanding color neutrality (c * = 1.33) and superior memory retention with only a 5% rise in transmittance after 350 min at 550 nm. Additionally, a large-scale (15 × 15 cm 2 ) solid-state REDM-based ECD was also fabricated to demonstrate the scalability of the approach. Moreover, a flexible REDM-based ECD has been realized using the organohydrogel electrolyte and annealed ITO/PEN sheets, which exhibits cycling stability for 400 cycles with an optical contrast of up to 43% and maintains excellent functionality after bending for up to 800 times. Our presented flexible solid-state ECD provides a possibility to enable the development of smart windows by mounting them onto existing plain glass or curved surfaces.
Preparation of Solid-State Gel Electrolytes: The preparation of traditional organogel electrolyte was referred to the previous work. 1.0 mmol NiCl 2 . 6H 2 O (237.69 mg), 0.5 mmol CuCl 2 (67.23 mg), and 4.5 mmol LiClO 4 (478.76 mg) were dissolved into DMSO (10 g) by magnetic stirring to obtain the liquid electrolyte. Then, 2 wt.% PVA powder (220.5 mg) was added into the previous solution and magnetically stirred in the oil bath at 70 °C until a homogeneous electrolyte was obtained. After cooling at room temperature, a traditional organogel electrolyte was obtained. In contrast, the solid-state gel electrolytes were prepared by freezing-thawing treatment of the corresponding gel electrolytes. Firstly, a liquid electrolyte was prepared similarly as above, except that mixed solvents (10 g) of DMSO and deionized water were used, with the DMSO fraction ranging from 100 to 60 wt.%. Then the liquid electrolyte was added with 8 wt.% PVA powder (0.94 g) and magnetically stirred at 90 °C until a homogeneous solution was formed. Subsequently, the cooled mixture was injected into a mold and vacuumed to eliminate bubbles within the solution. Finally, the mold was frozen at −20 °C for 20 h to facilitate the gelation process, followed by being thawed at room temperature to obtain the solid-state gel electrolyte. The gel electrolytes produced by the FT method are denoted as FT-DMSO x , depending on the weight percentage x (x = 100, 90, 80, 70, 60) of DMSO in the mixed solvents. Note that FT-DMSO 100 belongs to organogel electrolytes, while FT-DMSO 90/80/70/60 with water to organohydrogel electrolytes.
Fabrication of Solid-State Rigid and Flexible ECDs: For rigid ECDs, two FTO glasses (4.0 cm × 2.5 cm) acted as both working and counter electrodes. The Ni-Cu based solid-state organohydrogel electrolyte was sandwiched between two FTO glasses, and strap-shaped spacers (1 mm thickness) were attached on the four edges of the glasses to separate the electrodes and surround the electrolyte. The active area of the device was 2.5 cm × 1.5 cm. Flexible ECDs were assembled similarly, except that the two annealed ITO/PEN sheets was employed as working and counter electrodes.
Characterizations: The ionic conductivity of organohydrogels can be calculated according to Equation (1), where R denotes the resistance of the sample, L and A denotes the thickness of electrolyte membrane and contact surface between electrode and electrolyte, respectively. [35] R was obtained from the electrochemical impedance measurement, which was performed using Autolab PGSTAT30 potentiostat in the frequency range from 1 MHz to 0.01 Hz at an open-circuit potential.
The degree of crystallinity can be determined through DSC measurements. Samples weighing between 1 and 2 mg were obtained from the center of organohydrogels, temperature range was 50-250 °C with a heating rate of 10 °C min −1 . The degree of crystallinity (f c ) was calculated manually with Equation (2): where ΔH m and ΔH 0 m are the melting enthalpy of PVA in the gel electrolyte and pure PVA crystals (150 J g −1 ), [36] respectively.
The morphology of electrodeposited Ni-Cu alloy films on the working electrodes was observed by a field emission scanning electron microscope (FESEM 7600F) with Energy-dispersive X-ray spectroscopy (EDX) and atomic force microscope (AFM, Park Systems NX10). The crystallographic structure of the electrodeposited films and ITO/PEN sheets was obtained by an X-ray diffractometer (XRD, Bruker D8 Advance Powder). A UV-vis-NIR spectrometer (Cary 5000) was used to measure the transmittance spectra and kinetics spectra at 550 nm for the cyclic testing.

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