A Dual‐Function Micro‐Swiss‐Roll Device: High‐Power Supercapacitor and Biomolecule Probe

Tiny autonomous systems less than 1 mm across need small energy storage to satisfy the demand for temporary pulse power and consistent power supply. A battery or supercapacitor alone cannot meet both requirements. Integration of battery and supercapacitor is an alternative, but it introduces a device (supercapacitor) that is not frequently invoked but requires substantial space. Herein, a submillimeter (0.42 mm2) high‐power supercapacitor with an additional function, namely, an on‐chip integrated probe for biomolecule detection is created. The dual‐function device is directly used in the liquid electrolyte, and its Swiss‐roll geometry allows for operation in a minimum 12‐nL electrolyte. The micro‐Swiss‐roll supercapacitor delivers a power density of 911 mW cm−2 and displays a 98% capacitance retention over 10 000 cycles. The biomolecule probe achieves a sensitivity of 230–262 µA mm−1 with a limit of detection of 0.4–0.5 × 10−3 m for the proof‐of‐concept target, the neurotransmitter dopamine. The biomolecule probe achieves a 230–262 µA mm−1 sensitivity with a limit of detection of 0.4–0.5 × 10−3 m for the proof‐of‐concept target, the neurotransmitter dopamine, along with selectivity in the presence of ascorbic acid. The independent dual function provides a promising route toward the full‐time use of dust‐sized supercapacitors integrated into submillimeter functional systems.


Introduction
Microsystems have shrunk to the submillimeter regime, advancing the development of implantable technology, smart dust, and small autonomous systems. [1][2][3] At the submillimeter scale, energy storage that enables autonomous operation has emerged as a crucial limitation. [4] The challenges are manifold. It is fiddly to introduce wet chemistry (electrode slurries and electrolyte fluids) used to create batteries and supercapacitors (SCs) into fabrication procedures for microsystems. [5] Moreover, the stored energy decreases dramatically as the size reduces, making energy storage devices still a challenge for microrobots and microsystems. [6][7][8] A notable recent achievement is the use of a 3D selfassembly process at the microscale, also known as micro-origami, to self-fold thin-film batteries into micro-Swiss-roll devices less than one square millimeter across but that offer high energy density. [4,9,10] However, the lack of solid or quasi-solid electrolytes still denies the integration of micro-Swiss-roll energy storage devices into tiny systems. Addressing this challenge requires multidisciplinary efforts coming transversely from materials science, [11] electrochemistry, [12] and microelectronics. [4] One approach to deal with this challenge is to miniaturize sensors and microrobots, such as swimming microrobots, which therefore allows for the mobility in physiologically relevant fluids. [13] These application scenarios open up the possibility of using fluid media as the electrolyte, bypassing the challenge of solid-like electrolyte integration. A nanobiosupercapacitor operating in blood was reported very recently, demonstrating the feasibility of this concept. [14] The operation of microsystems also imposes challenges. Switches and actuators require pulsed temporary energy (shortterm energy supply) while other components may need a smooth and constant energy supply. Neither SCs nor batteries by themselves can satisfy both needs. An intuitive way to address this issue is to integrate both SCs and batteries. [15] In this scenario, the pulsed energy demand can be buffered by the SCs, whereas the batteries provide the system with stable energy. However, adding components to each other often leads to the increased Figure 1. Self-encapsulated SC based on Swiss-roll nanomembranes. a) Flowchart for the SC fabrication and self-rolling (namely, Swiss-roll packaging): a.1) sacrificial layer patterning on the glass substrate; a.2) hydrogel layer patterning; a.3) polyimide layer patterning; a.4) patterning of the current collectors (namely, interdigitated electrodes) and subsequent PEDOT coating via electrodeposition, whereas the PEDOT structure is illustrated in the inset; a.5) interdigitated SC and its optical microscopy image (top panel); a.6) sacrificial layer etching and SC self-packaging, as shown by the micrography at the top panel; and a.7) encapsulated SC and its optical microscopy image (top panel). Microscopy scale bars (a.5-a.7): 500 μm. b) Sketch of the ion transport process that occurs in the SC through the gap between and around the interdigitated electrodes. c) Sketch of the ion transport observed within the Swiss-roll device after its self-packaging. d) Height profiles of a polyimide substrate, Au current collector, and PEDOT thin-film coated Au current collector.
technological complexity of the integration. Furthermore, SCs are often invoked when a pulse power is required and return to standby mode, resulting in a "dead" component for the microsystem. Therefore, improving function density by taking advantage of the system's built-in electrochemistry, especially in SC architectures, is a straightforward way to enhance the limited space utilization of tiny systems. As the electrochemical processes are not limited to energy storage but are applied to actuators and sensors, [16,17] the multifunctional designs offer plenty of opportunities to microscopically exploit the features of both electrochemical energy conversion and storage. [18,19] Here, we demonstrate a strain-induced self-assembly technique to create a micro-Swiss-roll on a chip (0.42 mm 2 ) with dual function: supercapacitor and biomolecule probe. The micro-Swiss-roll device shows high power densities and long cycle life as a supercapacitor and good sensitivity and selectivity for biomolecules. The supercapacitor and biomolecule probe are independent of each other. Nevertheless, the dual-function Swissroll device is not an autonomous system. A trifling volume, viz., 12 nL, of 1 × 10 −3 m sulfuric acid (H 2 SO 4 ) solution is enough for the operation as an SC using poly(3,4-ethylenedioxy-thiophene) (PEDOT) as the electrode material. The micro-Swiss-roll SC provides up to 0.8 V with a high-power density (up to 911 mW cm −2 ). In addition, the micro-Swiss-roll SC displays outstanding capacitance retention (up to 98%) over 10 000 charge-discharge cycles. Finally, the redox reaction on the PEDOT film allows the selective detection of biomolecules such as dopamine (DA) in the presence of ascorbic acid (AA). [20] The micro-Swiss-roll device is capable of monitoring the DA concentration with well-defined sensitivity (230-262 μA mm −1 ) and limit of detection (0.4-0.5 × 10 −3 m). These findings provide a promising route toward the creation of dust-sized systems to convert and store electrochemical energy. The use of liquid electrolytes brings the further possibility of simultaneously exploiting redox electrochemical processes to develop multifunctional modules in microsystems.

Results and Discussion
The devices were fabricated by standard photolithography processes using a three-polymer layer stack (viz., polymeric sacrificial layer, hydrogel layer, and polyimide layer). [21] Figure 1 illustrates the fabrication steps of 3D self-assembled Swiss-roll devices. Figure 1a.1 depicts the sacrificial layer patterning on the glass substrate. Figure 1a.2 illustrates the patterning of the hydrogel layer on the top of the sacrificial layer. Afterward, the polyimide layer is spin-coated and subsequently patterned on the entire surface of the hydrogel layer (Figure 1a.3). Later, the gold (Au) interdigitated current collectors were deposited and coated with the PEDOT thin film electrode (Figure 1a.4). At this fabrication stage, the planar device architecture was obtained as shown in Figure 1a.5. The top panel in Figure 1a.5 is an optical microscopy image of the planar interdigitated SC. Furthermore, 3.7% hydrochloric acid (HCl) solution (pH = 1) was employed for the chemical etching of the sacrificial layer and to swell the hydrogel layer as well (Figure 1a The planar architecture can be considered an open electrochemical system for the ion transport processes (Figure 1b), while the Swiss roll encapsulates the ions and allows their migration and diffusion to occur only between the tube windings ( Figure 1c). Accordingly, the formation of the electrical double layer in planar devices takes place within a semispherical free space. [22] In contrast, the Swiss-roll device's electrical double layer is formed within a very low operating volume, ≈12 nL, provided by the tiny gap remaining in the interlayer spaces of the micro-Swiss roll structure (the volumes of hydrogel, polyimide, current collector, and PEDOT are discounted, and they are estimated considering the thicknesses of the dry layers). The thicknesses of the active layers of the SCs are visualized in Figure 1d, which exhibits the height profile of the polyimide substrate, the current collector (60 ± 3 nm thick film), and the PEDOT electrode (575 ± 38 nm thick film). Figure 2a shows the interdigitated current collector on the polymeric layers. In the magnified image, the geometric standards of our interdigitated devices are labeled: the length of each electrode finger is 2,000 μm, the width is 20 μm, and the gap between two adjacent electrodes is 20 μm. A total of 54 finger electrodes are created. Figure 2b illustrates the typical dimensions of the micro-Swiss-roll device. The external diameter of the ≈8winding micro-Swiss-roll is (166 ± 10) μm, whereas its internal diameter is (85 ± 7) μm. As the micro-Swiss-roll curvature is considerably larger than both the width and separation of the interdigitated electrodes, the slightly bent electrodes and current collectors of the first winding would resemble planar devices. Accordingly, to succeed with the complete device packaging into the Swiss-roll structure, the first 1-2 windings of the micro-Swiss-roll devices were designed to not consist of interdigitated electrodes. This was done by the introduction of a 620 μm long electrode-free space on the polymeric layer stack, as zoomed in Figure 2a.
The footprint area (FA) of the planar device was estimated considering the PEDOT-coated area as FA planar = (3.9 × 10 −2 ) cm 2 . For the micro-Swiss-roll devices, FA is the product between the length of the device that contains PEDOT (0.25 cm) and the external diameter, i.e., FA Swiss roll = (4.2 × 10 −3 ) cm 2 , a value ten times reduced in comparison to the planar device. Furthermore, the active volumes of planar and Swiss-roll devices were estimated as the product between the area of the PEDOT-coated surface and the thickness of the PEDOT thin film. Accordingly, the active volumes are the same for both device architectures, (2.2 × 10 −6 ) cm 3 . Thus, the Swiss-roll SC displays a ≈2.2 nL active volume and is expected to operate with a very small amount of electrolyte solution, ≈12 nL. In addition, the total volume of the devices was estimated by including both active and non-active materials (i.e., glass slide for the planar architecture, and hydrogel and polyimide layers for the Swiss-roll device). [1] The planar device shows a (3.7 × 10 −3 ) cm 3 total volume. The total volume of the Swiss-roll device is (5.8 × 10 −5 ) cm 3 , which corresponds to 0.42 % of the total volume of the planar SC. Thus, the 3D self-assembly process shrunk the total volume of the device by ≈240 times.
Two electrode layouts are used to investigate the electrolyte confinement effect on SC performance: finger electrodes longitudinal (l) and transverse (t) to the micro-Swiss-roll axis. The responses for cyclic voltammetry measurements (CV, Figure 2c-f) show near rectangular and symmetrical shapes for all four types of devices. Cyclic voltammograms exhibiting the statistical errors are available in Figure S1 (Supporting Information). In particular, micro-Swiss-roll SCs exhibit slightly-less square shapes, implying limitations imposed by the electrolyte confinement. [23,24] For the l electrode layout, individual straight fingers are displaced along different windings of the micro-Swiss roll. In this scenario, the winding separator, i.e., the polymeric layer stack, would suggest that the electrochemical performance rate may reach a limit in a short time. However, the results ( Figure 2d) are contrary to this expectation: the rate performance is almost identical to the t electrode layout, in which all the electrodes are displaced as individual Swiss-roll fingers in parallel to each other (zoomin, Figure 2f). Thereby the performance is independent of the electrode layout, confirming that major electrical changes do not arise from the different curvature conformation of finger electrodes on the interlayers of Swiss rolls. Such a feature provides the interdigitated-electrode Swiss-roll SCs with design flexibility for a multifunctional application.
The CV evaluation as a function of scan rate ( ) (Figure 2c-f) allowed us to calculate the capacitance (C) of each device at low , from 0.03 to 0.09 V s −1 . Figure 3a shows the SC capacitive current plotted as a function of , where C can be calculated from the slope of the linear regression of such a curve. [25] For the l layout (Figure 3a.1), we found C low planar, l = (29.3 ± 0.2) μF, and C low Swiss roll, l = (38.5 ± 0.1) μF. For the t layout (Figure 3a.2), C low planar, t = (37.1 ± 0.2) μF and C low Swiss roll, t = (41.1 ± 0.1) μF. The slightly higher capacitance found for the micro-Swiss-roll devices can be attributed to the electrolyte confinement. As previously reported, Swiss-roll electrodes lead to higher electric fields across the confined electrolyte, which improves ion migration rates and therefore increases the doublelayer capacitance. [12,16] As the electric double-layer formation process is dynamic, high-measurements can affect C. This can be verified by performing CV at 10 V s −1 (Figure 2c-f). From such CV curves, C can be calculated considering the discharge process and applying Equation (S1) (Supporting Information). [23] We found for the planar devices, C = 10 V/s planar, l = (28 ± 4) μF and C = 10 V/s planar, t = (36 ± 3) μF, whereas for Swiss-roll devices, C = 10 V/s Swiss roll, l = C = 10 V/s Swiss roll, t = (30 ± 2) μF. Electrochemical impedance spectroscopy was also performed. Figure 3b exhibits the impedance data acquired for the l and t devices, respectively. For better visualization of the data at different frequencies (Figure 3b), the complex planes were plotted on a logarithmic scale. In addition, the frequency variation from 0.1 to 10 5 Hz is depicted by the shaded background according to the right-hand-side color scale. [26] The impedance plots using linear scales are provided in Figure S2 (Supporting Information). Up to ≈80 Hz, the devices displayed a capacitive behavior, and it is noteworthy that the formation of micro-Swiss-roll devices does not affect the capacitive frequency regime. Capacitance values estimated from the impedance data are in Figure S3 (Supporting Information), plotted as a function of resistance (Z′) and frequency. Figure 3c shows impedance and phase variations as a function of frequency for l-and t-devices (panels c.1 and c.2, respectively). For phases below −45°(i.e., down to ≈−90°), the devices displayed capacitive behavior, which occurred up to ≈80 Hz. The equivalent series resistance, R ES , can be estimated by extrapolating the complex plane curve at higher frequencies. R ES is equivalent to the value of the real impedance Z′ when the imaginary impedance is equal to zero (Z″ = 0). [22,27] The R ES values for l electrode layout are R ES planar, l = (36 ± 3) Ω and R ES Swiss roll, l = (42 ± 2) Ω. For the t electrode design, R ES planar, t = (31 ± 3) Ω and R ES Swiss roll, t = (50 ± 2) Ω. The slight difference between R ES Swiss roll, l , and R ES Swiss roll, t is attributed to the spiral shapes of t interdigitated electrodes (zoom-in, Figure 2f). Considering the t layout, the Swiss-roll structure is more prone to impose small changes in the PEDOT thin-film conformation compared to the l layout, which explains such an increment observed in the t electrode series resistance after device self-rolling.
Galvanostatic charge-discharge (GCD) measurements were employed as an alternative approach to evaluating the devices' capacitance. For the GCD experiments, we used a 10 μA current ( Figure S4, Supporting Information). The C values for the l electrode layout (estimated according to Equation (S2), Supporting Information) were C planar, l = (44 ± 3) μF and C Swiss roll, l = (49 ± 3) μF. For the t devices, we found C planar, t = (38 ± 6) μF and C Swiss roll, t = (50 ± 1) μF. The areal capacitance (C FA ) was calculated as the ratio between the C values obtained from the GCD curves and FA. Microscale devices often impose stringent limits on the component footprint, making area properties better indicators of device performance than volumetric or gravimetric properties. [1,28,29] For the planar architectures, C FA planar, l = (1.1 ± 0.1) mF cm −2 and C FA planar, t = (0.8 ± 0.1) mF cm −2 . For Swiss-roll devices, C FA Swiss roll, l = C FA Swiss roll, t = (12.0 ± 0.6) mF cm −2 . In addition, the time constant can be estimated according to = R ES × C. [27] We found planar, l = (1.6 ± 0.2) ms, Swiss roll, l = (2.0 ± 0.3) ms, planar, t = (1.2 ± 0.2) ms, and Swiss roll, t = (2.5 ± 0.3) ms. On the one hand, the smaller values found for the planar devices indicate the confinement of electrolyte slightly reduces the capacitor performance. On the other hand, a higher as provided by the micro-Swiss roll implies a slower self-discharge. [1] In addition to C, the other two important parameters required for practical applications are the areal energy density (E FA ) and the areal power density (P FA ), Equations (S3) and (S4) (Supporting Information), respectively. [27] The planar devices exhibited E FA planar, l = (97 ± 9) nW h cm −2 and E FA planar, t = (71 ± 7) nW h cm −2 . The Swiss-roll architectures showed E FA Swiss roll, l = E FA Swiss roll, t = (3,330 ± 160) nW h cm −2 . Regarding P FA , we found P FA planar, l = (114 ± 6) mW cm −2 , P FA Swiss roll, l = (754 ± 30) mW cm −2 , P FA planar, t = (132 ± 7) mW cm −2 , and P FA Swiss roll, t = (911 ± 50) mW cm −2 . Figure 3d shows the capacitance retention over 10 000 chargedischarge cycles. The micro-Swiss-roll devices were submitted to conditioning cycles for 20 min before the 10 000-cycle measure to ensure reproducible electrolyte-interlayer penetration. The l-and t-planar devices exhibited capacitance retention of (86 ± 1)% and (89 ± 1)%, respectively (Figure 3d.1,d.3). The Swiss-roll devices exhibited excellent cycling stability with capacitance retentions of (95 ± 3)% and (98 ± 4)% for l and t electrodes, respectively (Figure 3d.2,d.4). It is noteworthy that the PEDOT film deposited on the planar device (i.e., an open system) is more subject to environmental conditions than in the Swiss roll. Therefore, the improved capacitance retention of the micro-Swiss-roll SC can be attributed to the confinement of the active materials, which restricts the PE-DOT surface to interact only with a small volume of electrolyte, and can avoid at some level the occurrence of side reactions. Oxygen and hydrogen evolution are examples of side reactions and can occur even with the cell operating within the potential window of thermodynamic stability of the electrolyte. [30,31] The side reactions can lead to energy losses, self-discharge of cells, and insufficient Coulombic efficiency. [30] As such, in Figure 2c-f, the slight current increment observed from 0.75 to 0.80 V may be ascribed to the beginning of oxygen evolution in the SCs. [23,31] For comparison, the state-of-the-art studies based on PEDOT Swissroll SCs show capacitance retention of ≈97% over 5000 chargedischarge cycles and ≈92% over 10,000 cycles. [32,33] The micro-Swiss-roll displayed capacitance retention as high as the stateof-the-art SCs, i.e., up to 99% over 5000 charge-discharge cycles and 98% over 10 000 cycles. Our findings show that the increase in C FA , E FA , and P FA , and cycling stability can be achieved consistently by confining the active area and the electrolyte within the micro-Swiss-roll windings. The performance comparison of the l-and t-SCs with recent devices is listed in Table S1 (Supporting Information). [18,[32][33][34][35][36][37][38][39][40][41][42][43] It is obvious that the Swiss-roll devices exhibited outstanding areal power density and cycling stability. Beyond previous micro-Swiss-roll SCs reported in the literature that include the electrodes in the initial windings without any confinement, [14,32,33] here we have quantified the performance of the fully confined system. In addition, our findings demonstrate that device performance is independent of the electrode layout, providing practical means to push forward the long-term capacitance retention in a microsystem.
In addition to SC measurements, the l-and t-Swiss-roll designs were investigated for biomolecule concentration monitoring as a proof-of-concept, aiming to extend the usability of such a Swiss-roll platform for the field of microsensors. [44,45] The test target was the biomolecule DA, a key neurotransmitter in the human central nervous, blood-vascular, urinary, and endocrine systems. [20] DA is also expected to function as a biomarker for diseases such as Parkinson's and Alzheimer's, depression, schizophrenia, and cardiotoxicities. [18,20] The DA detection mechanism using the micro-Swiss-roll devices is sketched in Figure  4a.1, where the l-and t-Swiss-roll device layouts are shown by the optical microscopy images in Figure 4a.2 and Figure 4a.3, respectively. When DA reaches the PEDOT electrodes, the biomolecule will lose electrons and be oxidized to dopamine-o-quinone. [46] The transferred electrons are transported toward the device current collectors via the PEDOT thin-film by means of subsequent charge-transfer reactions. [47] As such, the PEDOT electrode is shared by both the SC and the biomolecule-responsive probe, providing insight into the further development of dual-function devices. In addition, the DA oxidation process practically does not affect the SC operation ( Figure S5, Supporting Information).
Differential pulse voltammetry is used to detect DA in a concentration range from 0.01 to 6 × 10 −3 m (Figure 4b,c). A 20 minconditioning protocol was employed between each measurement to guarantee the complete penetration of the DA solution in the micro-Swiss-roll devices and therefore the effective contact of the biomolecule with the PEDOT electrodes. In addition to the conditioning time, the pH of the DA solution was adjusted to ≈2 to prevent further changes in the hydrogel layer conformation and avoid the spontaneous oxidation of DA before each 3) the t electrodes. b,c) Differential pulse voltammetry was measured using the l-and t-electrode designs, respectively. The DA-concentration range varied from zero (i.e., blank solution) to 6 × 10 −3 m, and the pH was from 2 to 3. The oxidation process resulted in the peak currents, as highlighted in orange (from 0.4 to 0.5 V). d,e) Calibration curves respectively obtained from the l-and t-electrode peak currents as a function of DA concentration. measurement. It is worth mentioning that DA detection is also possible using a pH of 6-7 as shown in Figure S6 (Supporting Information), making the device closer to real applications. However, we found that, at pH ≈ 2, the level of swelling of the hydrogel layer resulted in a suitable interlayer space as one may notice in the pronounced oxidation signatures (Figure 4b,c). [18] The DA oxidation process on the PEDOT electrodes provided well-defined peak currents around potential values from 0.4 to 0.5 V for both devices, l-and t-layouts, as shown by the highlighted regions in Figure 4b,c. Figure 4d,e shows the calibration curves obtained from the differential pulse voltammograms. For the two micro-Swiss-roll devices, the DA concentration monitoring performance is similar, which is consistent with the results that the electrode layout has little effect on the ion-transport path as previously discussed for the SCs. The sensitivity (S) for each electrode design is estimated from the slope of the linear regression of the calibration curves. [12] Two linear regions are found for each calibration plot, which results in two different values of S according to the concentration range. The l electrode exhibits S (1) l = (230 ± 9) μA mm −1 , and the t one has S (1) t = (262 ± 36) μA mm −1 . Furthermore, the limit of detection (LOD) for both designs was calculated as LOD = 3 × SD × S (1) −1 , where SD is the standard deviation of the linear regression. [12] We found LOD l = (0.5 ± 0.1) × 10 −3 m and LOD t = (0.4 ± 0.2) × 10 −3 m. Although the current LOD values are not sufficient for the use of the devices for DA detection in clinical samples (nanomolarto micromolar-concentration ranges are necessary for DA detection in blood plasma, urine, or brain), [48,49] future optimizations in the methodology (e.g., enzyme immobilization, pH adjustment, increase in the electrochemically active area) [20,46] are feasible to significantly improve the sensing capability in future Swiss-roll applications. In addition, to investigate the selectivity of the Swiss-roll device to DA, which is relevant for real applications, differential pulse voltammetry was performed in a solution containing 2.5 × 10 −3 m AA as the DA concentration changed ( Figure S7, Supporting Information). Accordingly, the PEDOTcoated Au electrode was selective for DA in the presence of AA.

Conclusion
PEDOT-based micro-Swiss-roll devices were successfully fabricated by standard photolithography processes with two electrode designs: longitudinal and transverse to the Swiss-roll axis. The 3D self-assembly process was proved to be an efficient miniaturization approach, creating a submillimeter (0.42 mm 2 ) device. The total volume (including inactive materials) of the devices is also reduced by up to 240 times compared to their planar counterparts. The micro-Swiss-roll SC shows excellent C FA , E FA , and P FA : C FA attains up to 12 mF cm −2 , E FA reaches 3.3 μW h cm −2 , and P FA is 911 mW cm −2 . It is noteworthy that the confinement of active materials and electrolytes avoids side reactions and results in stable cycling performance, retaining up to 98% of capacitance over 10 000 cycles. In addition, the control of the electrochemical processes in the micro-Swiss-roll device makes a dual-function device possible. The micro-Swiss-roll devices were employed for dopamine detection, displaying a sensitivity of 262 μA mm −1 and a limit of detection of 0.4 × 10 −3 m. The micro-Swiss-roll device also demonstrates the independence of electrode layouts, offer-ing good design flexibility toward multifunctional applications. The dual-function micro-Swiss-roll device suggests a paradigm to integrate energy storage in microsystems and offers a spaceefficient way to use microscale devices.
Preparation of Three-Layer Polymeric Stack-Based Devices: The threelayer polymeric stack was prepared sequentially with a ≈200-nm-thick sacrificial layer, a ≈1400-nm-thick hydrogel layer, and a ≈500-nm-thick polyimide layer. The thickness of each polymeric layer was measured in a dry state right after its patterning and baking process. The thicknesses were confirmed after device fabrication and Swiss-roll assembly by using control samples submitted to equivalent heat treatments. Height profiles for the thermally treated hydrogel layer before and after immersing the sample for 3 h in a 3.7% HCl solution were provided in Figure S8 (Supporting Information). For the patterning process, the sacrificial layer solution was deposited by spin-coating at 3000 rpm for 60 s on a clean glass substrate followed by a soft bake at 35°C for 3 min. The sample then was exposed to 350 nm-ultraviolet (UV) radiation using a Karl Süss MA6 Mask Aligner. The exposed patterns were developed in deionized water, dried, and then rinsed in propylene glycol monomethyl ether acetate (PGMEA). In addition, the samples were hard-baked at 220°C for 15 min. Afterward, the hydrogel layer solution was spin-coated at 2000 rpm for 45 s, and exposed to UV using the MA6 Mask Aligner. The developing process was performed in diethylene glycol monoethyl ether (DEGMEE) followed by rinsing in PG-MEA. Then, the samples were hard-baked at 220°C for 15 min. After the hydrogel layer, the polyimide solution was deposited on the substrate by spin-coating at 6000 rpm for 60 s. The samples were then exposed to UV and further developed in a solution containing ethanol, DEGMEE, and Nethyl-2-pyrrolidone (1:2:4 parts per volume). After exposure, a hard bake procedure was carried out at 220°C for 15 min. Subsequently, the interdigitated electrodes were patterned by a standard photolithography process using a negative AZ 5214E photoresist (MicroChemicals) and a Heidelberg MLA100 Maskless Aligner. Afterward, 10 nm of chromium (Cr) and 50 nm of Au were sequentially evaporated using an electron beam. In addition, a 5-μm-thick SU-8 3005 film was deposited on the contacts to boundary the active area of the electrodes.
After the microfabrication processes, PEDOT material was electrodeposited on the electrodes using chronoamperometry. The electrodeposition was carried out using the interdigitated electrodes as the working electrode, a silver/silver chloride reference electrode, and a platinum counter electrode. The chronoamperograms were acquired by applying 0.9 V for 2 min, using a solution containing 1 m H 2 SO 4 , 10 × 10 −3 m sodium dodecyl sulfate, and 10 × 10 −3 m 3,4-ethylenedioxythiophene. [32] Finally, the release of the hydrogel and polyimide layers from the substrate occurs with the chemical etching of the sacrificial in a 3.7% HCl solution (pH = 1). It is noteworthy that for the fabrication of the planar devices, just the polyimide layer and the electrodes were patterned on the glass substrates.
Electrochemical Measurements: Chronoamperometry, CV, and electrochemical impedance spectroscopy were performed using an Autolab PGSTAT204 potentiostat (Metrohm). For electrochemical characterization, ≈170 μL of electrolyte was added with a micropipette to a mold of polydimethylsiloxane (PDMS) sealed to the chip ( Figure S9, Supporting Information). The electrolyte penetrates the Swiss-roll architecture through its edges (Movie S1 and Figure S10, Supporting Information). For CV and electrochemical impedance spectroscopy, the electrolyte was 1 × 10 −3 m H 2 SO 4 . For the CV, the -range was from 0.01 to 10 V s −1 , with a potential window from 0 to 0.8 V. The GCD measurement was carried out using BioLogic Instrument. The operating voltage varied from 0 to 0.8 V, and the applied currents for charge and discharge cycles were 10 and −10 μA, respectively.
DA concentration Monitoring: Differential pulse voltammetry was performed using a μAutolab Type III. A two-electrode setup consisting of PEDOT-coated Au electrodes was employed for the analysis. DA solution was prepared using a 0.1 m HCl-potassium chloride (KCl) solution with pH 2 (blank solution). The DA concentration varied from 0.1 to 6 × 10 −3 m. After adding the DA solution, a conditioning time of 20 min was required to ensure that the solution penetrated between the windings of the Swiss roll. The parameters for the measurements were: operating voltage from 0 to 0.9 V, step potential of 0.01 V, modulation amplitude of 0.025 V, and modulation time of 0.2504 s.

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