Inkjet Printing Bio‐Inspired Electrochromic Pixels

In this report the design, fabrication, and testing of inkjet‐printed electrochromic pixels (ECPs) incorporating the biochrome, xanthommatin (Xa) as programmable display units is described. As a redox sensitive chromophore, Xa is present in some species as a physiological indicator with red (reduced) or yellow (oxidized) colors associated with different behavioral or developmental stages. These features have been recently leveraged in some materials applications, illustrating a bio‐inspired design solution to color‐changing sensors and displays. This paper describes an extension of these applications to print individually addressable ECPs that can be processed in a mild annealing step to introduce localized conductivity on initially nonconductive substrates. When formulated together with a poly(3,4‐ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) carrier ink, an addition of 0.19 wt% Xa is enough to generate dynamic ECPs which can be batch printed as lateral electrodes on any substrate to serve as both conductors and display units across electrically isolated boundaries. Application of low potentials triggers reversible color changes that span the red/yellow color space and can cycle for days. These results represent an important step towards the incorporation of alternative active materials like Xa to manufacture and scale low‐power, color‐changing pixels and patterns.


Inkjet Printing Bio-Inspired Electrochromic Pixels
Patrick A. Sullivan, Daniel J. Wilson activated events that alter visible colors. [4] However, these generalizations become far more complicated when considering full system integration, which requires the inclusion of controls that also distribute signals (mechanical, biochemical, and/ or electrical) across large surface areas for global color changes. Cephalopods (cuttlefish, octopus, and squid) are an example of a living system that can achieve this function by controlling nodes of densely populated dermal pigmented organs called chromatophores in less than a second to obscure or reveal light. [5] As such, they represent a powerful model for dynamic colorchanging systems.
Biomimetic systems designed to approximate chromatophore structure/ function properties have been iterated over the past decade [6] with more recent strategies based on mechanochromism, [7] magnetochromism, [8] thermochromism, [9] and photochromism. [10] Many of these technologies have demonstrated impressive strategies to control shape, color, and in some cases, patterns built to mimic the high speed, cooperative display changes in cephalopods. However, they remain challenged by requirements for complex control hardware, electrically expensive actuation strategies, and/or unsustainable materials that limit their use for many practical applications.
One strategy to overcome some of the intensive power requirements and throughput challenges associated with many of the cephalopod mimetic or inspired technologies is the use of electrochromic materials. Electrochromic devices (ECDs) rely on a voltage-activated electrochemical reaction that modulates the transmission of light and/or color value, typically in the visible color spectrum. Given that these chemical reactions typically require low (<5 V) voltages and sometimes require no continuous application of current to maintain a given transmissivity, they can result in ultra-low powered optical devices. [11] "Ultra-low" power for display technology is typically defined as <10 mW cm −2 , [12] but ECDs can frequently maintain a colored state with near-zero power consumption and operate at <1 mW cm −2 when switching states. While there is a variety of different types of electrochromic materials including metal oxides, [13] small molecules, [14] and conductive π-conjugated polymers, [15] conductive polymers have emerged as a promising, cost-effective material amenable to multimaterial integration and manufacturing. [16] There are currently multiple ways of preparing electrochromic devices with spin coating, [17] screen printing, [18] In this report the design, fabrication, and testing of inkjet-printed electrochromic pixels (ECPs) incorporating the biochrome, xanthommatin (Xa) as programmable display units is described. As a redox sensitive chromophore, Xa is present in some species as a physiological indicator with red (reduced) or yellow (oxidized) colors associated with different behavioral or developmental stages. These features have been recently leveraged in some materials applications, illustrating a bio-inspired design solution to color-changing sensors and displays. This paper describes an extension of these applications to print individually addressable ECPs that can be processed in a mild annealing step to introduce localized conductivity on initially nonconductive substrates. When formulated together with a poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) carrier ink, an addition of 0.19 wt% Xa is enough to generate dynamic ECPs which can be batch printed as lateral electrodes on any substrate to serve as both conductors and display units across electrically isolated boundaries. Application of low potentials triggers reversible color changes that span the red/yellow color space and can cycle for days. These results represent an important step towards the incorporation of alternative active materials like Xa to manufacture and scale low-power, color-changing pixels and patterns.

Introduction
Adaptive coloration, or camouflage, in nature is often used as a strategy for organisms to visually blend in with [1] or stand out from [2] their surroundings for defensive, communicative, or predatory purposes. [3] While the mechanisms regulating camouflage in nature are species-dependent, in most cases they can be generalized as mechanically, biochemically, and/or electrically electrodeposition, [19] and spray coating [20] being the most commonly used. However, these top-down methods require a significant amount of starting material for successful casting, much of which is applied in excess and often discarded as waste. An alternative to these methods is piezoelectric inkjet printing (IJP), which allows for the precise, and reproducible, deposition of ink from the bottom up-a process that minimizes waste by eliminating the need for masking or aligning steps for drop-on-demand deposition and subsequent patterning. [21] With piezoelectric IJP specifically, the voltage and frequency of each waveform can be manipulated to deform the piezoelectric printhead to expel a droplet from each nozzle (between 1 and 12 nozzles). [21e,g] Because of the flexibility of IJP, generally any aqueous ink with viscosities ranging from 8 to 40 mPas and surface tensions of 40 mN m −1 can be optimized for printing through a user-controlled waveform. [22] Recent work has seen the development of inkjet-printed electrochromic devices utilizing a conductive π-conjugated polymer, poly(3,4-ethylenedi oxythiophene):poly(styrenesulfonate) (PEDOT:PSS), for wearable display, electrode array, and sensor applications. [21b,c,23] PEDOT:PSS is a commonly used π-conjugated polymer in generating electrochromic materials because of its ability to form a dispersion in water, tunable conductivity, biocompatibility, and highly reversible transparent (oxidized) to dark blue (reduced) color change. [24] We asked whether the adaptive features of inkjet printing could be leveraged together with the biochrome xanthommatin (Xa), commonly found in the skin of cephalopods, to create electrochromic pixels (ECPs) with colors that expand beyond the blue/ clear states of traditional PEDOT:PSS in a fabrication scheme that allows for the production of low-power displays. As a redoxactive chromophore, [25] Xa can undergo a voltage-triggered color change from yellow (oxidized) to red (reduced), which we have recently demonstrated in manually prepared devices templated on PEDOT:PSS-coated indium-tin-oxide (ITO) substrates. [19] We observed that the PEDOT:PSS matrix offers two advantages: conductivity and charge transfer to facilitate color change, while Xa, introduced at significantly lower weight percentages than the matrix, dominates the visible color changes. Despite the device performance, these devices were not easily scalable, limiting their practical utility as pixels in dynamic displays.
In this report, we describe the production and application of inkjet-printed, Xa-based ECPs, where we demonstrate the ability to scale and control an array of pixels. In our design, each ECP is built laterally to minimize the number of postprint processing and assembly steps needed to build and array functional devices. These printed electrodes are spatially separated to serve as both conductors and display elements, illustrating a pathway for batch-printing and controlling electroactive, colorchanging systems using accessible electronics that enable both local and global changes in color.

Xanthommatin Synthesis
Prior to printing, a fresh ink containing newly synthesized Xa was formulated. [26] Briefly, a 3-hydroxykynurenine (4 mg, 17.8 µmol) (3OHK) was dissolved in sodium hydroxide (25 µm, 1 mL) and mixed dropwise with a solution of potassium ferricyanide (16.5 g, 50.1 µmol) dissolved in deionized water (0.5 mL) at room temperature for 90 min. The resulting Xa was precipitated by addition of 1 m hydrochloric acid (0.5 mL) followed by treatment with DI water over multiple centrifugation and wash cycles (Eppendorf Centrifuge 5254).

Printing Lateral Electrochromic Pixels
A piezoelectric inkjet printer (Dimatix Materials Printer, DMP-2850) was used for all printing. Inks were formulated with freshly synthesized Xa dissolved in a pH 6, 0.050 m MES (2-(N-morpholino)ethanesulfonic acid) buffer mixed with PEDOT-Jet40 ink (Pedotinks.com) to produce a final concentration ranging from 0-3 mg mL −1 Xa. These inks were then filtered through a 1-µm polytetrafluoroethylene syringe filter (Thermofisher) before loading into a Samba print cartridge chamber (Fujifilm) compatible with the DMP-2850. The cartridges were degassed in a vacuum chamber for 30 minutes. Prior to printing, 0.1 mm thick sheets of polyethylene terephthalate (PET) were washed, dried, and treated with a UV-ozone cleaner (Jetlight, Model 30) for 30 minutes. Printing was carried out at room temperature, 22 °C, where 1-5 nozzles were used consistently with a 33 kHz jetting frequency at 39 V. Following printing, the patterns were annealed at 60 °C. For the full-scale device each pattern consisted of two spatially separated electrodes with the same area (1.7 cm 2 ). The gap between each electrode was 1 mm, and each had 1 cm leads extending from their edges to facilitate connecting to the external power supply. This was uniformly scaled down to yield 12.5% sized devices compared to the full-scale device.

Assembling the Electrochromic Pixels
After printing and annealing, the outer borders of the ECPs were sealed with a thin polydimethylsiloxane (PDMS) gasket (Sylgard 184 Silicone Elastomer Kit) to create an open-faced chamber for the electrolyte during testing. UV-curable glue (Bondic) was then applied to the edge of this gasket to prevent leakage of the liquid electrolyte (propylene carbonate (2 g), 1-butyl-3-methylimidazolium tetrafluoroborate (1.6 g), and hydroquinone (0.003 g)).

Calculating Z-Value
All inks were characterized for printability using the reciprocal of the Ohnesorge number (Z value, Equation 1). [27] Z d ( ) Here, the density (ρ) was determined by measuring the mass of 100 µL of the ink with an analytical balance (VWR-225AC). Viscosity (η) was measured using a Discovery Hybrid Rheometer-2 (TA instruments) using a standard-size 40-mm 1° cone (TA instruments) and plate geometry. These measured www.advmatinterfaces.de viscosities were derived under a ramp shear rate of 1-100 s -1 over 5 minutes. Surface tension (γ) was measured with an Attension Theta optical tensiometer (Attension Instruments). Finally, the drop diameter (d) was calculated based on the assumption that a 2.4 pL spherical drop (diameter of 1.66 µm) is produced from each nozzle during printing. All inks were characterized with n = 3 measurements per sample.

Characterization of Printed Patterns
Conductivity and resistance were measured with an Ossila 4-point probe. During testing, 1 cm 2 square patterns were analyzed, making three measurements per square pattern across three independent patterns. Thickness was measured similarly (three measurements over three samples) using a Dektak 3030CR Profilometer. All static patterns were imaged using an Epson Perfection V19 scanner with a resolution of 300 dpi in picture color mode with a white background.

Electrochemical Characterization
Chronoamperometry profiles of the ECPs was collected using a Gamry Interface1000B Potentiostat under +/-0.5, 1.5 and 2.5 V vs the reference electrode in a two-electrode setup with the reference electrode shorted to the counter electrode for 150 seconds each. Each device ran for five cycles prior to data collection (Supporting Information Figure S1). Data was collected every second. All experiments were performed across three independent pixels. The current was recorded every second for the duration of the experiment. For longer term experiments (>10 cycles), data points were collected every 10 seconds. Chronopotentiometry data was collected across three pixels every 0.1 second at 50, 100, 200 and 400 µA.

Colorimetric Analysis
Videos of the ECPs were captured using either an iPhone 8 plus (12-megapixel resolution camera, 60 frames per second, with Adobe Premiere Rush as the video collection application) or a time-lapse camera (Brinno TLC2020). The time-lapse camera was used exclusively for studies >10 cycles, wherein a single frame was captured every 10 sec. In each case, these videos were deconstructed into a single photo every 10 (<10 cycles) or 30 seconds (>10 cycles) with a video editing software (Videoproc Converter). The images were then compiled into an image sequence that was analyzed with ImageJ [28] and a custom-built macro, which measured associated RGB values of the inner and outer electrodes over the course of the image sequence with RGB Measure, an ImageJ plugin (https://imagej.nih.gov/ij/ plugins/rgb-measure.html).

Control Strategy for ECP Arrays
To control the printed 4-pixel arrays, a microcontroller-based power supply was used to control the redox states of the pixels individually. This system used a standard microcontroller (Arduino Uno) to control the magnitude and polarity of voltages directed through each ECP using H-bridge breakout boards (SparkFun Electronics), enabling control of the oxidation states of electrochromic blends in timed sequences. +/−1.5 V was applied across each ECP to induce redox changes, and the cycles lasted for 300 seconds.

Reflectance Measurements
Reflectance spectra were collected with a Flame Miniature Spectrometer (Ocean Insight) with a halogen lamp light source (Ocean Insight). Each measurement was calibrated against a Spectralon diffuse reflectance standard. The reflectance probe was positioned over the inner electrode of the 100% scale ECP at an angle of 45°, and the spectra was collected every 15 seconds as the ECP was cycled at +/− 0.5, 1.5 and 2.5 V vs the reference electrode in a two-electrode setup for 150 seconds each.

Calculating Color from Reflectance Data
To better correlate the reflectance spectra with the visible colors observed during ECP cycling, a MATLAB script [29] was used to convert all measured values to the International Commission on Illumination (CIE) xyY color space. Briefly, this script integrates the reflectance measurements of each sample normalized by the relative power of the illuminant over the 380-780 nm range with respect to the normalized spectral sensitivity of human cone cells. Following this abstraction, the brightness of the color can be deconvoluted from its chromaticity, which is expressed as Cartesian coordinates in a CIE color space.

Color Efficiency
Color efficiency was calculated by measuring the change in optical density (ΔO.D.) of the reduced and oxidized states at 555 nm as a function of the change in charge density per area. [30] The % reflectance at 555 nm was collected over 10 cycles, and the R ox and R Red are the average % reflectance across those 10 cycles recorded at the end of the reduction or oxidation half cycle. Here, ΔO.D. is measured according to Equation 2.

Results and Discussion
In formulating our Xa-based inks, we first investigated the role of Xa concentration on printability, conductivity, and color. To start, we incorporated up to 3 mg mL −1 Xa into a PEDOT:PSS vehicle ink. We then characterized the fluid physical properties of the modified inks using a theoretical Z-value calculation that relates viscous forces to inertial forces and the solution surface tension needed for inkjet printing (Equation 1). [27b] The Z-values www.advmatinterfaces.de derived for inks formulated with 0, 1, 2, and 3 mg mL −1 Xa were 6.56 ± 0.66, 6.80 ± 0.48, 8.57 ± 0.06, and 9.04 ± 1.04, respectively ( Figure 1A). While we noticed that the addition of Xa increased the Z-values due to its increased density contributions, all values fell between the specified 1-10 range recommended for successful printing.
[27a] We next printed inks as simple squares and characterized all patterns on a nonconductive PET substrate ( Figure 1B), selected intentionally to eliminate the need for conductive substrates like indium tin oxide (ITO) coated glass or plastic. Because the lateral ECP design requires electrically isolated anodes and cathodes, we leveraged the natural conductivity of annealed PEDOT:PSS to serve this function. To achieve this, the devices were cured overnight in a 60 °C oven, a process that facilitated complete evaporation of the liquid phase from the printed patterns to increase conductivity ( Figure S2, Supporting Information). To optimize the ECPs, we first varied the number of printed layers (1, 3, 5, 7, and 10) and characterized the thickness and conductivity of each. We observed thicknesses ranging from 0.108 ± 0.015 µm (1 printed layer) to 1.031 ± 0.025 µm (10 printed layers) ( Figure 1C). We next used these thickness values together with each pattern's surface resistivity to calculate sample conductivity. We observed that conductivity increased as the printed layer count increased across all modified inks and that the amount of Xa doped into the polymer matrix directly impacted conductivity ( Figure 1D). The higher the Xa concentration, the lower the conductivity ( Figure S3, Supporting Information). This difference is likely a result of higher defects in the annealed polymer network, which decreases entanglement and conductivity compared to the control (no Xa).
We next examined whether these blended patterns could still activate color change in a lateral pixel configuration. To test this, we patterned pixels as two adjacent electrodes spatially separated on a continuous substrate, where the printed inner and outer electrodes had the same surface area with different geometries-an important design feature for balancing charge transfer in this device (Figure 2A). In this configuration, we positioned a flexible PDMS gasket over each pixel to contain an ionic liquid electrolyte and connected the device in series with an external power supply to control the voltage at each electrode. Color and color change across the electrodes were characterized through pixel intensity measurements, where individual RGB values associated with each electrode over 10 cycles were recorded to understand color development upon application of a voltage. Upon analysis of the RGB color channels, we observed the largest difference in signal in the green channel ( Figure S4, Supporting Information), which was then used to compare across multiple devices. It is important to note here that RGB analysis as a metric is inherently highly dependent on relative imaging conditions, making it difficult to compare raw values across multiple devices. Thus, we report the change in the green channel (ΔGreen) when comparing performance across different devices. Using this metric, we analyzed the effect of Xa concentration on color development in our 7-layer system. We observed a ΔGreen value of 9.3 ± 3.4 in the inner electrode of devices printed with 1 mg mL −1 Xa, despite its relatively high conductivity. At 3 mg mL −1 Xa, we observed selective areas of more vibrant colors from Xa, even though it had the lowest overall conductivity, but saw poor overall color change across the device with no discernible www.advmatinterfaces.de cycles, making it impossible to analyze the color change in a meaningful way. The 2 mg mL −1 Xa condition yielded the most consistent colors that were preserved over multiple cycling events with the largest ΔGreen of 19.5 ± 0.9 ( Figure S5, Supporting Information). We next investigated the effect of print layers on color development. We observed the emergence of visibly contrasting colors in the 7-and 10-layer devices with ΔGreen values of 19.5 ± 0.9 and 24.9 ± 0.4 respectively. Visible colors observed from devices printed with five layers were challenging to differentiate, producing ΔGreen values of 5.8 ± 2.8 ( Figure S6, Supporting Information).
Moving forward, we employed the 10-layer 2 mg mL −1 Xa formulation to build and test the ECPs, highlighting that only 0.19 wt.% Xa in the ink is needed to generate the dynamic color changes observed in the devices. In these conditions, we observed reproducible cycling between the reduced and oxidized states for both the outer and inner electrodes with minimal decay in the green color intensity over 10 cycles ( Figure 2B). Chronoamperometry was used together with video imaging to understand color evolution and reproducibility over time ( Figure 2C, Figure S7, Supporting Information.) Maximal color contrast appeared within 150 s, which is 15x longer than our PEDOT:PSS control patterns (Videos S1 and S2, Supporting Information). This time delay is due to the material differences in the electrodes with and without Xa. These differences are even more pronounced in our ECPs, which are printed as lateral electrodes, activated at one fixed point on the periphery of the pattern. For the electrodes to trigger color change, the charge must move from the activation point across the entirety of the electrode area -a process regulated by surface area and conductivity, amplified more so considering the absence of a conductive substrate like ITO to facilitate electron transfer. Together, these interfacial differences highlight an important trade-off that should be considered in applications of our inkjetprinted ECPs. Despite its slower activation, the ECPs demonstrated a sustained function over 500 cycles equating to 41 h and 40 min of use (5-minute cycle times) with minimal degradation in color or performance ( Figure 2D, Figure S8, Supporting Information).
We next asked whether it would be possible to expand and selectively control an array of ECPs. To test this, we patterned four pixels using a slightly modified design that enabled access to the leads via an off-board microcontroller ( Figure 2E). The printing was carried out in a single batch from one pattern, highlighting the ability to develop full displays of increasing complexity through facile pattern development without the need for postprocessing assembly. By creating timed sequences of control signals delivered to each pixel, we developed a series of programs that highlight their potential for display applications, where each pixel is defined by the color state of the inner electrode. In the first demonstration, we performed a simple on/off program, wherein the device cycles between all four pixels being reduced or oxidized with the application of a +/−1.5 V (Video S3, Supporting Information). We expanded this Figure 2. Single pixel design and characterization. Ai) Schematic of the device assembly. Aii,Aiii) show top view schematics of the device (left) with a potential of −1.5 V and +1.5 V, respectively. Scale bars correspond to 1 cm. Both schematics have an associated photo of the state in a physical 10-layer 2 mgmL -1 Xa device next to them (right). B) Normalized Green RGB value tracked over 10 cycles of device function alternating between +1.5 V and −1.5 V every 150 s. C) Chronoamperometry results show the current across the device during cycling as described above. D) Color analysis of the first and last 5 cycles of our ECPs function over a 500-cycle test period. Error bars in B and C represent standard deviation of n = 3 replicates per condition. E) Demonstration of display capabilities of the printed ECP arrays using a "chase" program, with the associated photos of each stage of the program. Red dashed circle highlights the pixel undergoing reduction.
www.advmatinterfaces.de demonstration in a chase program, wherein a single electrode is reduced and the other three are oxidized clockwise, resulting in a pattern of all four pixels cycling around a wheel over the course of 20-minutes ( Figure 2E, Video S4, Supporting Information). When taken together, this data illustrates the ability to program color changes in multiple, electrically isolated ECPs on one substrate produced from a single print pattern.
To better understand the operating conditions needed for optimal ECP performance, we investigated the effect of the applied voltage and current on color generation. Using chronoamperometric analysis, we observed that the application of potentials ranging from +/−0.5-2.5 V all generated similar plateau currents of 0.05 ± 0.02 mA (reduced) and −0.03 ± 0.01 mA (oxidized) but that the time-to-plateau decreased with increasing voltages (Figure 3A). This trend also extended to color change, which was recorded when the measured current value reached 95% of its total change. Specifically, at +/− 0.5, 1.5, and 2.5 V the device takes 115.7 ± 14.1, 79.7 ± 7.4, and 48.7 ± 11.6 s, respectively, to display discernable differences in color. These findings were complemented by RGB analysis, where ΔGreen values increased from 17.4 ± 0.1, 24.9 ± 0.4, and 33.5 ± 0.1 for 0.5, 1.5, and 2.5 V, respectively (Figure 3Bi,Bii). Reflectance spectra were also collected for the inner electrode using a 45° diffuse reflectance probe, where we observed wavelength shifts from 620 to 595 nm along with intensity shifts at 420 nm corresponding to the reduced and oxidized states of the inner electrode only ( Figure S9, Supporting Information). Because of our extensive RGB analysis, we chose to focus these experiments on the inner electrode only, as it provided the largest continuous surface area for spectral analysis without compromising the signal-to-noise when compared to data collected along the thinner, outer electrode. In our experiments, we observed % reflectance changes for the oxidized and reduced states recorded at 420 nm of 20.0 ± 0.8% and 40.9 ± 1.6% for 0.5 V, 28.8 ± 3.6% and 48.9 ± 0.5% for 1.5 V, and 35.7 ± 2.2% and 43.2 ± 2.4% for 2.5 V, respectively. These spectral data were then transformed using a MATLAB script [29] to provide CIE color space coordinates and enable an output of computer-generated color swatches that matched visible colors generated in our ECPs ( Figure S9, Supporting Information). This analysis highlighted that despite the fairly minor changes observed in the total visible reflectance spectra, these values do indeed correlate to visible color differences in the red/yellow color spaces.
From here, we next determined the color efficiency (CE) at 555 nm of each device operating at the different potentials, where 555 nm was selected to represent maximal visual contrast for humans ( Table 1). [31] We observed a CE of 62.8, 71.9, and 74.1 cm 2 C −1 when +/−0.5, 1.5, and 2.5 V were applied, respectively, further supporting that while the ECPs still function at lower voltages, maximal contrast was achieved at potentials >0.5 V. At voltages >2.5 V, the color changes in the ECPs were inconsistent and in some cases irreversible, suggesting that the material may be transformed under these conditions.
To better understand the mechanisms driving this voltagedependent color change, we conducted chronopotentiometry experiments under galvanostatic conditions, where each pixel  was held at 50, 100, 200, and 400 µA and changes in voltage were tracked over time ( Figure 3C; Table S2, Supporting Information). These upper and lower bounds were selected based on the data presented in Figure 3A. Here, each experiment was run until the recorded voltage reached a vertical asymptote, indicating that the reduced or oxidized species in the electrode have been depleted. At the lower bound (50 µA), the inner electrode reached this point at 203.5 ± 5.6 seconds with a final operating voltage of 1.7 ± 0.2 V during reduction and 200.8 ± 1.2 s at −2.2 ± 0.1 V during oxidation. At the upper bound (400 µA), the device reached this point at 7.2 ± 0.2 seconds with a final operating voltage of 3.4 ± 0.2 V during reduction and 7.7 ± 7.0 seconds at −6.2 ± 1.9 V during oxidation. Despite this significantly faster switching speed, we again observed limited color change across the entire ECP at the higher currents/voltages ( Figure S10, Supporting Information). Instead, the color change was localized to the edges of the pixels closest to the leads. These data suggest that current-controlled experiments allow for faster switching than the voltage-controlled experiments from Figure 3 but that they do so at a cost of color quality and efficiency. At higher current values, the charge does not distribute effectively throughout the electrode, limiting its ability to generate visible color changes. This limitation may be due to the lower inherent conductivity of the printed electrodes, which are operating in the absence of external conductive substrates. At lower currents, ECPs are afforded the time to distribute charge across the entire electrode and, by extension, full-color development. These data suggest that while the onset of the voltage transition and subsequent color change are dependent on current magnitude and time, both variables are dependent on the printed geometry of the ECP itself.
To further explore this hypothesis, we miniaturized the surface area of our electrodes and tracked performance changes under the application of different voltages. To test this, we reduced our original pixel size by 1/8 to produce an internal electrode diameter of 0.375 cm (Figure 4A). When activated at +/−1.5 V over 10 cycles, the miniaturized devices generated ΔGreen values of 38.3 ± 5.4, which is 68% higher than the original devices (Supporting Information Figure S11A). When activated at 0.5, 1.5, and 2.5 V, these devices also generate color ≈4x faster (43.3 ± 4.1, 25.7 ± 3.3, and 19.0 ± 1.6 s, respectively) than the original devices ( Figure S11, Supporting Information; Video S5, Supporting Information). Calculated charge densities were also consistently higher at all applied voltages for the smaller devices with the largest jump from 0.5 to 1.5 V and no significant difference in 1.5 and 2.5 V ( Figure 4B). These data suggest that the higher charge densities afforded by the smaller devices are sufficient for faster visible color changes.

Conclusion
In the quest to develop accessible and adaptive optical technologies, we describe the design, fabrication, and characterization of inkjet-printed electrochromic pixels, which are generalized as electrically isolated electrodes spatially patterned onto nonconductive substrates using inkjet printing. While printing electrochromic materials including metallo-supramolecular polymers, [33] metal oxides, [34] and conductive polymers, [23a,35] has been done, we report the first formulations of Xa-containing inks and show how small (≈0.19 wt.%) amounts material can maintain electrochemical activity while dominating the visible (red/yellow) colors of each pixel at low voltages. Before considering such display applications, we must first move away from liquid-based electrolytes that require dedicated gasket enclosures and manual fluid handling to enable performance. One solution could be a transition to gel-based or solid-state electrolytes as a promising next step that could make the device more deployable. Such electrolytes could easily be removed and replaced to help increase the lifetime of the device. To support its deployment to multiple environments (moving and static structures), other future work will include onboarding all electronics and power supplies to further aid in device handling. While further development is necessary, our results demonstrate a promising first step toward producing scalable, reproducible, and precise electrochromic patterns that could be miniaturized to produce sustainable low-power displays inspired by natural systems.

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