*3.1. Materials*

Monomer 3,4-ethylenedioxythiophene (EDOT, 97.0% purity), κ-carrageenan (κC, ACS reagent), potassium chloride (KCl, >99.0% purity), dexamethasone 21-phosphate disodium salt (Dx, 98.0% purity), ammonium acetate (NH4CH3CO2, 98.0% purity), ultrapure water MS quality, and MS methanol were purchased from Sigma Aldrich (San José, Costa Rica). All chemical reagents were used without further purification.

### *3.2. Synthesis and Preparation of the Modified PEDOT:*κ*C:Dx Electrode*

Electrodes (20.49 ± 0.02 mm2) were fabricated by the deposition of gold on a polyimide substrate (see Figure S4) and they were passivated using a shadow mask to leave a specific exposed area to the electrode [55]. Prior to the polymer deposition, all electrodes were electrochemically cleaned applying cyclic voltammetry (CV) sweeps from a range of −600 to 900 mV with 100 mV·s<sup>−</sup><sup>1</sup> scan rate, in KCl 0.2 M [56], using an Autolab Potentiostat supplied by Metrohm (PGSTAT-302N, AUTOLAB, Utrecht, The Netherlands).

The surfactant dispersion was prepared according to a previous work [30], briefly: κC (0.2% *w*/*v*) and KCl (0.2 M) were added to deionized water previously heated at 50 ◦C. The samples were sonicated using 140 Joules in a Sonifier QSonica (Q700, Ultrasonic Corporation, Danbury, CT, USA), before and after adding the monomer EDOT (10 mM) and Dx at three di fferent concentrations: 1 mM, 5 mM, and 10 mM.

The solution was electropolymerized on the electrode surface using galvanostatic conditions in the Autolab Potentiostat. The gold electrode (see Figure S4) is used as working electrode, platinum as counter electrode, and Ag|AgCl (KCl 3.0 M) works as reference electrode. The electrical polymerization was carried out with a constant current of 102.45 microamperes (current density: 0.5 mA·cm<sup>−</sup>2) using a potential limit of 1400 mV during 360 s (ca. 180 mC·cm<sup>−</sup><sup>2</sup> of charge density). Following the PEDOT:κC:Dx deposition, the electrodes were intensively rinsed with deionized water and stored at 4 ◦C before their use.

### *3.3. Evaluation of the Stability and Size of the Dispersion Systems*

The characterization of the particle size and ζ-potential was performed using six dispersions, prepared in deionized water, namely: (1) κC 0.2% *w*/*v*; (2) Dx 10 mM; (3) EDOT 10 mM:κC 0.2% *w*/*v*; (4) EDOT 10 mM:Dx 10 mM; (5) κC 0.2% *w*/*v*:Dx 10 mM; and (6) EDOT 10 mM:κC 0.2% *w*/*v*:Dx 10 mM. Measurements were performed in a Zetasizer instrument (Nano ZS, Malvern Panalytical Ltd., Worcestershire, UK) at 25 ◦C and 173◦ angle. All the measurements were done by triplicate. Finally, dispersions were sonicated using a high-power ultrasonic bath (Bransonic ®, Merck corporation, San José, Costa Rica) for 6 min to promote their homogenization. Two more formulations of EDOT:κC:Dx were prepared to reach lower dexamethasone concentrations into the conductive layer.

### *3.4. Analysis of the Topography and Composition of PEDOT:*κ*C:Dx Coating by Profilometry and* μ*-Raman Spectroscopy Methods*

The electrode topography was studied by profilometry analysis (Bruker, model: Dektak TX Advance, AZ, USA) and the arithmetical mean roughness of the surface (Sa) was calculated to describe the topography of the materials by using a 2 μm tip radius and a force of 1 mg in a 300 × 300 μm<sup>2</sup> and a scan area rate of 2.5 <sup>μ</sup>m·s<sup>−</sup>1.

Raman spectroscopy analysis was carried out using a confocal μ-Raman microscope (Alpha300 R WITec, GmbH, Ulm, Germany) with a 532nm excitation laser, exposure time of 0.5 s, and 105 accumulations. The Raman stack scan was obtained using an integration time of 4 s in 4 μm<sup>2</sup> of area, 200 measurements per line were recorded for a total of 20 lines in each stack. Oversampling was used to improve the image quality, which was done in case of the cross-sectional scan. The scan depth was fixed at 5 μm and a total of 10 stack scans were achieved. The intensity of the relative wavenumber at 1435 cm<sup>−</sup><sup>1</sup> and 1625 cm<sup>−</sup><sup>1</sup> were extracted from each acquired spectrum, corresponding to PEDOT [2] and Dx/κC [44,57], respectively and plotted as 2D image. The intensity counts are related to the presence of the functional group and it is presented as bright yellow areas.

### *3.5. Dexamethasone Release Experiments from the PEDOT:*κ*C:Dx Film*

The Dx release from the modified electrode was carried out in a continuous flow cell using cyclic voltammetry (CV) sweeps with a three electrodes system (PEDOT:κC:Dx, Ag|AgCl and a gold film as working, reference, and counter electrodes, respectively). The active release of the drug was performed in 1 mL of fresh ammonium acetate solution (0.10 M) pH 7.2 [58], by scanning of CV from −600 to 1000 mV with a 25 mV·s<sup>−</sup><sup>1</sup> scan rate, over a period of 300 min (5 samples total) at room temperature.

The second release event, without electrical stimulation, was performed in order to analyze and to quantify the passive drug release process. For the experiment, 1.0 mL of 0.10 M ammonium acetate was injected through the cell containing the electrodes, a total of five samples were collected during 300 min of analysis.

Dexamethasone phosphate concentration, in the samples for the active and passive release events, was determined using a Xevo G2-XS quadrupole time of flight (Q-tof) mass spectrometer (Waters Corporation, Wilmslow, UK) coupled with an Acquity UPLC H-Class. For the analysis, a 10-μL injection of the sample was separated with an Acquity UPLC ® C18 column (2.1 mm × 50.0 mm). The mobile phase consisted of a solution of water:formic acid 0.05% *v*/*v* and methanol:formic acid 0.05% *v*/*v* and they were supplied under not isocratic conditions with a constant flow of 0.3 mL·min−<sup>1</sup> (Table S1).

The mass spectrometer was configured according to the parameters in a previous work [59], with the modifications shown in supplementary information S1. Quantification was carried out using Multiple Reaction Monitoring (MRM) acquisition method with the optimized transition of 471.1584 *m*/*z* for the precursor ion and 78.9585 *m*/*z* for the product ion, with a collision energy of 35 eV. Concentration in each sample was calculated using the Software MassLynx ™ (V4.1, Waters Corporation, Wilmslow, UK) and an external calibration curve between 0.5 ppb to 5000 ppb of dexamethasone phosphate (R<sup>2</sup> = 0.9965).
