High voltage flexible glucose/O2 fully printed hydrogel-based enzymatic fuel cell

Herein we report on a novel enzymatic fuel cell (EFC) based on stencil printed electrodes modified with pyrrolo quinoline quinone glucose dehydrogenase and bilirubin oxidase, which are assembled by considering two different configurations: (i) normal assembling in liquid electrolyte and (ii) six EFCs connected in series, each one comprising both bioanode and biocathode, coupled through a hydrogel-based electrolyte in a stack-like mode similar to a Voltaic pile. After a deep electrodes characterization, they are assembled according to the first configuration obtaining an open circuit voltage (OCV) of 0.562 ± 0.002 V. Moreover, the EFC performance are substantially improved by using the second configuration (six EFCs connected in series) obtaining an OCV of 2.36 ± 0.22 V with a maximum power output of 22.9 ± 0.9 μW at a cell voltage of 1.95 V (operating in 10 mM D-glucose). This innovative approach represents a proof-of-concept towards the development of renewable power sources and could serve as a critical step in powering implantable bioelectronics, such as pacemakers.

Particularly, many examples of implanted EFCs attempting energy harnessing from blood or other bodily fluids (as source of glucose) have been investigated to power low energy demanding electronic devices (e.g.insulin pumps, pacemakers, capacitors etc) [3,[6][7][8].This represented a key achievement towards standalone devices that could revolutionize the development of Point-of-Care (PoC)/Point-of-Need (PoN) devices making them available in remote area of the World [9].However, since 1970s, it was clear that the power/energy density supplied by EFCs could not be comparable with traditional batteries or fuel cells (FCs), hence not able to power high energy demanding devices like smartphones or computers [10].Moreover, EFCs also addressed biotechnologist/biochemists towards enzyme engineering to enlarge the library of enzymes (wild type or mutant) able to perform/improve direct electron transfer (DET) between redox centers and electrodes, minimizing their O 2 responsiveness (O 2 is the natural electron acceptor in many biological processes) [4,11].Since PQQ-GDH exhibits reduced substrate specificity and decreased stability compared to GOx, many protein engineers focused on improving its properties towards biosensors and EFCs development [12].
There are also other technological constrains at EFCs development such as limited open-circuit voltage (OCV) and limited energy density supplied [3].In these regards, novel printing technologies (e.g.stencil-printing, screen-printing, ink-jet printing, 3D-printing, and laser printing etc) and series cells stacking can tackle these issues [13][14][15].In all printing technologies, ink formulation plays a key role particularly considering both rheological properties (viscosity, shearthinning) and electrochemical properties (electroactive area, roughness factor) [16][17][18][19].In addition, most of the aforementioned printing technologies require post-production modifications like atomic layer deposition or electrodeposition, electrochemical polishing, solvent activation, enzyme activation, thermal activation, etc, while stencil-printing offers a cost effective and easy end-user fabrication alternative without any post-production process considering the possibility of integrating enzymes already within the conductive ink [20][21][22].
Recently, there has been an increasing interest towards water-based conductive inks especially due to their lower cost, higher conductivity and room temperature curing procedures compared to traditional organic solvents-based conductive inks [23,24].The latter exhibit low conductivity and poison proteins, hence hindering their utilization for electrode manufacturing [19].Reproducibility and operational stability represent the challenge for the development of wearable technologies including biosensors and EFCs [9].In this regard, the enzymes integration within conductive inks can be a solution to improve reproducibility, operational stability, storage stability, stiffness, and robustness of electrodes for EFCs and biosensing [25,26].
For in vivo applications, sterilization, preventing blood clots (i.e.electrode treatment with heparin), biocompatibility and lifetime should be accounted.This will be a key step towards implantable biofuel cells [27].
Furthermore, hydrogel-based electrolytes have been considered for developing several electrochemical devices [28,29].Despite their high ionic conductivity, liquid electrolytes can be dangerous for batteries construction considering uneven spill of concentrated inorganic salts solution (acidic/basic electrolytes), spill of organic solvents, low electrochemical stabilities (limited voltage range due to H 2 O oxidation/reduction) and low miniaturization [30,31].All these issues hinder their possibility to power miniaturized/wearable bioelectronics, often in direct contact with bodily fluids.Conversely, hydrogel-based electrolytes, despite their lower ionic conductivity, can be exploited to develop flexible, stretchable, and wearable electronics (e.g.self-powered biosensors, biosupercapacitors, actuators etc) [32].
Moreover, hydrogel-based electrolytes offer enhanced safety and adaptability when compared to traditional liquid electrolytes in flexible energy storage devices.They also have the potential to enhance the electrochemical performance of these devices.In addition to these advantages, hydrogel electrolytes can be engineered to possess self-protective, thermotolerant, self-healing, and other desirable properties, contributing to their rapidly advancing development [28,33,34].
Nishizawa and colleagues, for instance, utilized an agarose hydrogel as both the electrolyte and fuel retainer in a flexible w-BFC (wearable biofuel cell).They combined a multi-walled carbon nanotube (MWCNTs)-modified carbon fiber (CF) anode with immobilized D-FDH (fructose dehydrogenase) and a MWCNT-CF cathode with immobilized BOD (BOx) to assemble the biofuel cell.Impressively, this configuration achieved a maximum power density of 550 µW cm -2 at 0.4 V, maintaining its performance even when the biofuel cell was bent to an angle of 44 degrees.
EFCs rely on redox enzymes that catalyze oxidation of electron donors and reduction of oxidants like molecular oxygen [26,35,36].Despite their wide employment both as biosensors or EFCs anode, oxidases do not represent the 'ideal' candidates for electrode construction because they are unable to transfer electrons directly to the electrode requiring natural or artificial electron acceptors/shuttles [37].In particular, O 2 is reduced to H 2 O 2 that can diffuse towards the cathodic electrode causing its poisoning and EFCs performance worsening.On the other hand, artificial electron acceptors/shuttles need to be properly selected to minimize the thermodynamic barrier of the electron transfer process between the enzymatic redox center and the electrodes [38,39].Pyrroloquinoline quinone (PQQ)-dependent glucose dehydrogenase (PQQ-GDH) represents a valid alternative being oxygen insensitive and able to transfer electrons directly to the electrode [12].Notably, BOx from Myrothecium verrucaria (MvBOx) has been widely employed as cathodic electrode to conceive EFCs working at human physiological pH.BOx catalyzes a 4e − /4H + reduction of O 2 to H 2 O with a low overpotential, being its formal potential (E 0′ ) close to the thermodynamic O 2 /H 2 O redox potential (i.e.+0.69 V vs. SCE) [40].
This work reports a novel fully printed EFCs with hydrogelbased electrolyte to maximize the OCV and power output (P out ) as reported in figure 1(A).Figure 1(B) shows schematically the water-based carbon ink formulation to print flexible electrodes by stencil-printing (SPG) and the connection in series of six EFCs (smaller than AAA battery) encompassing a bioanode based on PQQ-GDH catalyzing glucose oxidation and a biocathode based on MvBOx catalyzing O 2 reduction.Both the electrodes and the hydrogel were characterized through electrochemical measurements (electroactive area, electron transfer rate constant, electrochemical impedance spectroscopy (EIS)), spectroscopic and morphological measurements to optimize the EFCs assembling.Finally, the device was assembled and tested including in a model solution (10 mM D-glucose) within the hydrogel.

Electrode preparation and modification
The water-based conductive ink was formulated using graphite, chitosan, and glycerol [19].Graphite served as the conductive material, chitosan as the binder, and glycerol as the stabilizer.A 2.5% w/v chitosan solution was prepared by dissolving chitosan in 1 M acetic acid.The solution was stirred at room temperature overnight.Subsequently, the chitosan solution was diluted to a final concentration of 1% w/v with distilled water, resulting in an acetic acid concentration of 0.4 M.
To prepare the conductive ink, 5 g of graphite powder was mixed with 10 ml of the previously prepared chitosan solution, along with 500 µl of glycerol.
The SPG electrodes were prepared using polyethylene terephthalate (PET) sheets.The sheets were cleaned three times with IPA and distilled water, and then sanded with fine emery paper (1500 grit) to enhance the adhesion of the ink.A stencil was created on a Smart Vinyl adhesive sheet using a Circuit Explore ® 3 machine equipped with Design Space Software v.7.3.95.The stencil was applied onto the PET sheet, and 500 µl of the conductive ink was placed on the PET sheet and spread using a scrape.The prepared electrode was allowed to dry at room temperature for 10 min and then cured in an oven at 100 • C for 1 min.Afterward, the stencil was peeled off, and the connecting track between the working electrode and the pad was insulated using nail polish.

Hydrogel-based electrolyte
To prepare the hydrogel beads, sodium alginate was dissolved in water and stirred overnight at 30 • C until achieving a final concentration of 2% w/v [41,42].The alginate solution containing the protein was then added drop by drop to a 20 mM aqueous solution of CaCl 2 .As a result, millimetersized hydrogel beads were formed through the crosslinking reaction between the alginate and calcium ions present in the CaCl 2 solution.The beads were allowed to maturate for 15 min in the CaCl 2 solution, during which further crosslinking occurred and the beads achieved their final stability and shape.Finally, the beads were cut to achieve the disk shape with an optimized thickness of 2 mm, figure S1 (supporting information).The Ca 2+ -crosslinked alginate hydrogel layer was frozen in liquid N 2 and immediately dried under vacuum overnight.Imaging was performed in the center and the edger part of the layer.

EFC assembly
The EFC was assembled according to two different configurations for PQQ-GDH bioanode and BOx biocathode: (i) two stencil printed electrodes in liquid electrolyte; (ii) six EFCs assembled in series with hydrogel-based stacks embedded with the electrolyte, as shown in figures 1(C) and (D), respectively.Finally, the EFC setups were tested in a model solution (10 mM D-glucose in 10 mM HEPES buffer pH 7.5).

Electrochemical measurements
Cyclic voltammetry (CV), EIS and amperometry experiments were performed using a PalmSens4 electrochemical workstation equipped with PSTrace 5.6 v software.All potentials were measured using a BASi Ag|AgCl|KCl, 3 M (all potential values reported in the paper need to be considered toward this reference) and a platinum wire as reference and counter electrode, respectively.SPG electrodes (geometric area = 1 cm 2 ) were used as working electrodes.Polarization curves were recorded using linear sweep voltammetry in potentiostatic mode at a scan rate of 1 mV s −1 , ranging from the OCV to 0 V.The bioanode, PQQ-GDH/SPG, was employed as the working electrode, while BOx/SPG served as a combined reference and counter electrode.In the first assembly mode, both biocomponents were placed in an electrochemical cell without the use of a proton-exchange membrane.The experimental setup included 10 mM glucose in a 10 mM HEPES buffer at pH 7.5.
In the second assembly mode, the six EFCs were connected in series using hydrogel stacks.These hydrogel stacks contained 10 mM glucose in a 10 mM HEPES buffer at pH 7.5.The temperature-controlled experiments were conducted using a cryostatic bath (LAUDA RM6, Delran, NJ, USA) with a precision of ±0.01 • C.

Morphological and spectroscopic characterization
The morphological characterization was conducted using a field emission scanning electron microscope (FE-SEM) model ∑ igma Zeiss (Jena, Germany).Images were acquired in topview using the in-lens detector, with a 5 KV acceleration voltage, 4 mm working distance, 30 µm opening, and no additional sample treatment.
X-ray photoelectron spectroscopy (XPS) analyses were performed using a Versa Probe II Scanning XPS spectrometer from Physical Electronics GmbH.The instrument was equipped with an AlKα source, and the spot size was set to 200 µm.Wide-scan and high-resolution spectra were obtained in CAE mode, with a pass energy of 117.40 eV and 29.35 eV, respectively.The source power was set to 49.2 W. To compensate for sample charging, an electron gun operating at 1.0 V and 20.0 µA was utilized.The acquired data were analyzed using the MultiPak v. 9.9.0.8 software.

Electrochemical, spectroscopic and morphological characterization of printed electrodes
Graphite printed electrodes were prepared as previously described by stencil-printing of water-based graphite conductive ink onto flexible PET.
The electrodes were preliminarily characterized by CV in 5 mM Fe(CN) 6 3−/4− , at different scan rates (5-150 mV s −1 ), as shown in figure 2(A).SPG electrodes displayed a peakto-peak separation of 0.228 V at 50 mV s −1 and an electroactive area (A EA ) of 4.05 ± 0.4 cm 2 (calculated according to Randles-Ševc ˇík equation, figure 2(A) inset), hence a roughness factor (ρ) of 4.05 (considering a geometric area (A GEO ) = 1 cm 2 ).Moreover, the electron transfer rate efficiency has been evaluated accounting for a mixed model based on an extended method merging the Klingler-Kochi and Nicholson and Shain methods that could be applied either to completely irreversible or reversible systems [43].The electron transfer rate constant (k 0 ) was (9.3 ± 0.6) × 10 −3 cm s −1 .SPG electrodes were further characterized by using EIS in ferri/ferrocynide solution and applying a Ca 2+ cross-linked alginate hydrogel containing the redox probe to investigate the resistance to the charge transfer (R CT ), as reported in figure 2(B).In particular, SPG electrodes exhibited an R CT of (2.969 ± 0.041) × 10 3 Ω in ferri/ferrocynide solution, which is in good agreement with the literature [22], and R CT of (2.947 ± 0.536) × 10 4 Ω applying a Ca 2+ crosslinked alginate hydrogel containing ferri/ferrocynide solution.The R CT difference can be certainly ascribed to a different ionic conductivity between the electrolyte solution, typically 1.15 S cm −1 (10 mM HEPES buffer containing 100 mM KCl (as supporting electrolyte)), and 0.027 S cm −1 for Ca 2+ crosslinked alginate hydrogel, as reported in the literature [44].
The specific capacitance of the electrode at various scan rates were calculated basing on the CV curves according to the following equation: where I is the response current density (A g −1 ), V is the potential (V), v the potential scan rate (V s −1 ), and m is the mass of the active material in the electrode (g) [45].Figure 2(C) reports CV curves in 10 mM HEPES buffer pH 7.2 + 100 mM KCl at different scan rates, but the CV curve at 5 mV s −1 was used to calculate the specific capacitance, resulting to be 158 ± 7 F g −1 .The CV shape of bare SPG was nearly rectangular, which indicated that the sample had good conductivity and was close to ideal electric double-layer capacitors (EDLC).When the scan rate increased to 300 mV s −1 , the curve shape still maintained nearly rectangular, which indicated bare SPG played superior capacitive performance even at high current density and signified good electrochemical performance as an EDLC electrode material.The resulting EFC can behave as a supercapacitor being able to store the charge derived by oxidation/reduction process catalyzed by the enzymes immobilized onto the current collector.
The SPG electrodes characterized by SEM display a rough surface (figure 2(D)), confirming the high roughness factor of these electrodes.The surface chemical composition analysis of the electrodes was performed by using XPS to investigate the influence of curing time (0-180 s) at fixed temperature and temperature effect (room temperature-180 • C) at fixed curing time, as shown in figure 2(E).From the line shape of C1s spectral region it was possible to observe the overlapping of all the spectra, demonstrating that both the curing time and the temperature do not induce any oxidation/degradation of the electrodes surface.

Bioelectrocatalytic characterization of PQQ-GDH/SPG electrode
To assess the catalytic behavior of modified anode electrode, notably SPG with physisorbed PQQ-GDH, several CVs experiments in 10 mM HEPES buffer pH 7.5 + 100 mM NaClO 4 by adding 10 mM D-glucose as substrate.The same experiments were performed by using Ca 2+ cross-linked alginate hydrogel containing testing solution.
Figure 3(A) shows the CVs in the absence (black curve) and presence of 10 mM D-glucose (red curve).In nonturnover conditions, PQQ-GDH/SPG electrodes did not show any redox peak eventually correlated with non-turnover redox activity.After substrate addition, a catalytic curve starting at E ONSET = −0.065V with maximum current of 13 µA at E = 0.4 V (figure 3(A), red curve) was observed.Similarly, PQQ-GDH/SPG electrodes were tested by using The calibration curve for PQQ-GDH/SPG electrodes in liquid electrolyte (spanning overall the 1 × 10 −6 -2.5 × 10 −2 M), reported in figure 3(C), indicated a linear range from 10 µM to 2.5 mM (figure 3(C), inset), a detection limit of 3 ± 1 µM and a sensitivity of 1.7 µA mM −1 and a correlation coefficient of 0.99 (RSD 5.3%, n = 10).Additionally, the calibration curve was fitted to determine the classical Michaelis-Menten kinetic parameters, which resulted in I max of 8.5 ± 0.3 µA and an apparent Michaelis-Menten constant (K M app ) of 2.2 ± 0.2 mM (similar to K M measured in solution).The calibration curve for PQQ-GDH/SPG electrodes in hydrogel-based electrolyte reported in figure 3(D) (spanning overall the 1 × 10 −6 -2.5 × 10 −2 M), indicated a linear range from 25 µM to 2.5 mM (figure 3(D), inset), a detection limit of 10 ± 1 µM and sensitivity of 0.88 µA mM −1 and a correlation coefficient of 0.99 (RSD 5.4%, n = 10).The kinetic parameters were determined as I max of 8.1 ± 0.4 µA and a K M app of 4.5 ± 0.6 mM (higher K M than the ones measured in solution for homogeneous and heterogeneous catalysis).This can be ascribed to the porosity of hydrogel-based solid electrolyte, which affects the diffusion of the substrate towards the electrode.
The effects of pH and temperature on the proposed PQQ-GDH/SPG were evaluated in both cases, namely the liquid electrolyte and hydrogel-based electrolyte, and the results are shown in figures 3(E) and (F), respectively.Three different buffers have been tested, acetate, MOPS and HEPES buffer in order to cover the range of pHs between 3 and 10.The current signal increased with pH until 7.5, which corresponds to the maximum current signal obtained in HEPES buffer (figures 3(E) and (F)), by using either the liquid electrolyte or hydrogel-based electrolyte.10 mM HEPES buffer pH 7.5 is therefore chosen for our experiments as an experimental condition closer to human physiological conditions.The optimum temperature in liquid electrolyte resulted to be 35 • C, as shown in figure 3(E).These results are in close agreement with the optimum pH and T values reported in literature for PQQ-GDH [36].Similarly, the optimum temperature in hydrogel-based electrolyte resulted to be 30 • C, probably due to the instability of the hydrogel-based electrolyte at temperatures above that value [44].
The selectivity of PQQ-GDH/SPG electrode was studied in order to assess the influence of possible interfering compounds on its response.The signal obtained for a fixed concentration of D-glucose was compared to that obtained with equal amounts of interfering compounds such as D-fructose, D-galactose, D-mannitol, L-lactate, ascorbic acid, dopamine, and uric acid as shown in figures 3(G) and (H).No significant current responses were recorded except for ascorbic acid and uric acid, notably 10% and 12%, 8% and 6% of D-glucose signal in liquid and hydrogel-based electrolyte, respectively.In contrast to the other electrochemical interferents that were examined, ascorbic acid and uric acid display a notably higher diffusion coefficient (D = 5.9 × 10 −6 cm 2 s −1 for ascorbic acid and D = 7.5 × 10 −6 cm 2 s −1 for uric acid).Moreover, they exhibit easy oxidation at the electrode surface, specifically at the potential utilized for analytical measurements, whereas dopamine possesses a diffusion coefficient of D = 8.2 × 10 −7 cm 2 s −1 .The exceptional selectivity observed can be attributed to the relatively low operating potential of the third-generation electrode.
The storage stability of PQQ-GDH/SPG electrodes were investigated by monitoring the signal decrease of the current when the electrode is used for ten measurements every day over a period of 90 days for a 10 mM D-glucose solution, as reported in figures 3(G) and (H).The PQQ-GDH/SPG electrodes showed a signal decrease of its initial response of less than 10% after 90 days by using a liquid electrolyte, and less than 20% by using a hydrogel-based electrolyte.This can be probably due to a combination of the intrinsic stability of the enzyme immobilized onto SPG electrodes as well as the porosity and stability of hydrogel-based electrolyte.

Bioelectrocatalytic characterization of BOx/SPG electrode
The electrocatalytic behavior of BOx/SPG was investigated in 10 mM HEPES buffer pH 7.5 + 100 mM NaClO 4 , as liquid electrolyte and Ca 2+ cross-linked alginate hydrogel containing testing solution by scanning the electrode in anaerobic conditions (figure 4(A), black curve) and in equilibrium with air (figure 4(A), red curve), respectively.Figure 4(B) (black curve) shows a CV recorded in anaerobic conditions in solution (not possible to perform with the hydrogel-based electrolyte) and with the hydrogel system in equilibrium with air (figure 4(B), red curve).From the conducted experiments, distinct electrocatalytic waves were observed, which can be primarily attributed electrostatic orientation of BOx near to its hydrophobic region adjacent to the T1 site.This specific arrangement enhances the DET reaction, resulting in an amplified electrocatalytic response.The catalytic curve, reported in figure 4(B), is mass-transfer limited probably due to the porosity of the hydrogel-based solid electrolyte where O 2 has lower diffusion rate compared to the liquid electrolyte.
To prove unequivocally the DET mechanism, BOx/SPG electrodes were analyzed by amperometry in anaerobic conditions (removing oxygen by purging nitrogen in solution) and in equilibrium with air, showing a current decrease to 22 ± 1 µA, as reported in figure 4(C).This highlights the biocatalytic ORR occurring at the enzyme modified electrode.
Before assembling the EFC with the hydrogel-based electrolyte, both the anode and cathode electrodes, namely PQQ-GDH/SPG and BOx/SPG, were characterized by open circuit potentiometry (OCP) to assess their individual operational stability over 500 h, as shown in figure 4(D) (anode: red curve and cathode: black curve).In particular, PQQ-GDH/SPG electrodes showed an OCP of −0.068 ± 0.003 V and BOx/SPG electrodes showed an OCP of 0.502 ± 0.004 V, displaying an increased stability over 500 h, probably due to the porosity of the electrodes, which affects the amount of enzyme molecules immobilized and its stability.After assembling the EFC, the OCV resulted to be 0.562 ± 0.002 V, which is in good agreement with the OCV that can be calculated as a difference between the OCP curves measured individually, namely 0.570 V.

Characterization of EFC assembled in series by using hydrogel-based electrolyte
Finally, the EFC was assembled with the hydrogel-based electrolyte reporting an OCV resulted to be 0.560 ± 0.005 V, close to the theoretical difference between OCPs of bioanode and biocathode, figure 5(A).
A maximum power output of 5.9 ± 0.6 µW at a cell voltage of 0.54 V was obtained, figure 5(A).The stability of the EFC was further evaluated by running polarization curves every 15 days (for a total of 3 months) and measuring continuously the OCV over the same range of time, figures 5(B) and (C).After a month, the performance was decreasing in terms of both OCV (5%) and maximum power output (10%), respectively.The power output of a single EFC dropped to 2.4 ± 0.1 µW after 3 months.This behavior can be mainly attributed to the progressive hydrogel and biocatalysts degradation.Moreover, the flexibility of EFC was assessed by bending the electrodes up to 50 times (figure S2) observing a 75% (in solution) and 68% (in the hydrogel-based electrolyte) of power output retaining.
Figure 5(D) reports the power output curve for the EFC assembled in series, up to EFCs.Considering 6 EFCs assembled in series in a stack-structure (like Voltaic pile), exhibiting an OCV of 2.36 ± 0.22 V with a maximum power output of 22.9 ± 0.9 µW at a cell voltage of 1.95 V (operating in 10 mM D-glucose). Figure 5(D) reports also photos of Ca 2+ cross-linked alginate hydrogel stack and a SEM picture showing the porosity of the hydrogel layer.
A comparison of the results with those obtained by other researchers with EFCs based on GDH modified bioanodes and Myrothecium verrucaria BOx as biocathodes has been reported in table 1.
The power density output and OCV values obtained in our experiments consistently align with those reported in existing literature, indicating their comparability [46][47][48][49][50].However, it is worth noting that there is a notable exception: an electrochemical FC (EFC) utilizing the combination of FADGDH/menadione/MWCNTs || Lac/MWCNTs achieved a power output of 799.36 µW cm −2 .In our case, the relatively lower power output can be attributed to potential kinetic limitations in the interfacial electron transfer at both electrodes, as well as the stability of the enzyme layer.These factors likely contribute to the observed decrease in power generation efficiency.

Conclusions
In this study, we have successfully demonstrated the viability of efficient DET of PQQ-GDH and BOx enzymes on graphite stencil printed electrodes.We optimized and characterized two modified enzymatic electrodes to achieve optimal performance, which were then combined to assemble an EFC.Subsequently, to obtain an EFC with an OCV capable of powering small electronic devices like pacemakers or artificial pancreases, six EFCs were connected in series using hydrogel-based electrolyte stacks.
The assembled biodevice was tested in a model solution containing 10 mm glucose, yielding promising results.The OCV obtained was 2.36 ± 0.22 V, with a maximum power output of 22.9 ± 0.9 µW at a cell voltage of 1.95 V.The relatively low maximum power output was primarily influenced by factors such as low electroactive area (4.05 ± 0.4 cm 2 , which is typical for graphite electrodes), substrate diffusion through the hydrogel stack and the presence of various interferents, leading to biofouling of the electrochemical surface.Moreover, the EFC activity over time is strongly affected by the decreasing enzymatic activity (∼20% drop in 1 month of continuous operation).
To the best of our knowledge, this is the first report on the assembly of EFCs in series using stencil printed electrodes in combination with hydrogel stacks.This innovative approach represents a proof-of-concept towards the development of renewable power sources and could serve as a critical step in powering implantable bioelectronics, such as pacemakers.

Figure 1 .
Figure 1.(A) Scheme of the high voltage flexible glucose/O 2 fully printed hydrogel-based electrolyte enzymatic fuel cell.The electrodes are modified with PQQ-glucose dehydrogenase and bilirubin oxidase and stacked with Ca 2+ -alginate hydrogel (inset: egg-box structure); (B) the graphite modified electrodes are prepared by stencil-printing from water-based graphite conductive ink previously formulated; (C) configuration (i) two stencil printed electrodes and (D) configuration (ii) 6 EFCs assembled in series with hydrogel-based stacks embedded with the electrolyte.