Abstract
Progress towards continued integration of technology into living organisms requires electrical power sources that are biocompatible, mechanically compliant, and able to harness the chemical energy available inside biological systems. Conventional batteries were not designed with these criteria in mind. The electric organ of the knifefish Electrophorus electricus is, however, an example of an electrical power source that operates within biological constraints while featuring power characteristics that include peak potential differences of 600 V and currents of 1 A1,2. Here, we introduce an eel-inspired power concept that employs gradients of ions across miniature hydrogel compartments bounded by a repeating sequence of cation- and anion-selective hydrogel membranes. The system employs a scalable stacking or folding geometry that generates 110 V or 27 mW/m2 per gel cell upon simultaneous, self-registered mechanical contact activation of thousands of gel compartments in series; this design also circumvents power dissipation before contact. Unlike typical batteries, these systems are soft, flexible, transparent, and potentially biocompatible. These characteristics make artificial electric organs compelling for powering next generation implant materials in hybrids of living and nonliving systems3–6.
The ability to generate electrical discharges by excitable cells has evolved independently at least six times in natural history1,7. In particular, Electrophorus electricus, a knifefish commonly known as the “electric eel,” is a system optimized by natural selection for power generation from ionic gradients8,9; its specialized electric organs can generate discharges of 100 W10 entirely from the flux of small ions. The eel employs the resulting transient current spikes to defend itself as well as to detect and incapacitate prey2,8. From the point of view of engineering an electrical power source for operation within a living organism9, the electric organ of Electrophorus provides a fascinating example as it takes advantage of specialized anatomy and physiology (Fig. 1a,b) that employs thousands of membranes with densely-packed, exquisitely selective (Extended Data Table 1), and actuatable ion channels for generating large voltages and currents1,2.
Figure 1 |. Morphology and mechanism of action of the eel’s electric organ and the artificial electric organ.
a, Arrangement of electrocytes within the electric organs of Electrophorus electricus. Close-up shows ion fluxes in the firing state. b, Mechanism of voltage generation in electrocytes. Each cell’s posterior membrane is innervated and densely packed with voltage-gated Na+ channels; the anterior membrane is non-innervated and has papillar projections extending into the extracellular compartment that increase its surface area2. In the resting state, open K+ channels in both membranes produce equal and opposite transmembrane potentials of −85 mV, so the total transcellular potential is zero. During an impulse, the Na+ channels in the posterior membrane open and K+ channels close in response to neural signals, generating an action potential of +65 mV from the resulting change in relative permeability to Na+ and K+ ions (see Section S1) and a total transcellular potential across both membranes of +150 mV.2 c, Artificial electric organ in its printed implementation. In this and all subsequent figures, red hydrogel contains concentrated NaCl and was polymerized from neutral monomers, green gel was polymerized from negatively-charged monomers and is cation-selective, blue gel contains dilute NaCl and was polymerized from neutral monomers, and yellow gel was polymerized from positively-charged monomers and is anion-selective. d, Mechanism of voltage generation in artificial electric organ. Mechanical contact brings together a sequence of gels such that ionic gradients are formed across alternating charge-selective membranes, producing potentials across each membrane that add up as tetramers and can be stacked in series of thousands of gels (see Section S1).
Inspired by Electrophorus, we engineered a potentially biocompatible artificial electric organ by employing durable and accessible components as well as automated and scalable fabrication processes. This artificial electric organ is capable of generating potential differences in excess of 100 V by implementing three unique features of Electrophorus’ electric organs.
The first feature that evolved in the eel’s electric organ is the arrangement of thousands of ion gradients in series by growing long and thin electrically active cells known as electrocytes in parallel stacks that span the rear 80% of the eel’s body11 (Fig. 1a,b). The anterior and posterior membranes of electrocytes are bound, top and bottom, by insulating connective tissue and function as separate membranes with selectivity for two different ions such that the transcellular potentials across both membranes add up in series 1,2,12. Figure 1b illustrates the mechanism of action for generating potential differences and electrical discharges by a resting and firing electrocyte: In the resting state, the anterior and posterior membrane potentials cancel each other out. In contrast, during an impulse, the posterior membrane depolarizes to produce a total transcellular potential of approximately 150 mV (Fig. 1b). Large electric eels stack thousands of electrocytes in series and can generate potential differences over 600 V2; parallel arrangement of multiple stacks enables peak currents that approach 1 A at short circuit8,13.
To generate an artificial electric organ, we mimicked the anatomy of the eel by using four compositions of hydrogel as analogs of the four major components of an electrocyte, namely its two membranes with different ion selectivity on the anterior and posterior side, as well as its intracellular and extracellular salt compartments (Fig. 1b, Extended Data Fig. 1, Supplementary Information Section S1). Figure 1c,d illustrates that upon initiation of registered contact, tetrameric repeating units of a high salinity hydrogel, a cation-selective gel, a low salinity gel, and an anion-selective gel in sequence formed ionically conductive pathways, establishing electrolyte gradients across tens to thousands of permselective hydrogel compartments. Employing the principle of reverse electrodialysis14, each of these “tetrameric gel cells” generated 130–185 mV at open circuit (Fig. 1d), a value comparable to the potential generated by a single electrocyte2. Figure 2e,f illustrates that the potential differences arising from 2,449 gels stacked in series added linearly to reach 110 V.
Figure 2 |. Fluidic and printed artificial electric organs.
a, Left: Cartoon of a fluidic artificial electric organ before and after contact activation. Aqueous plugs of hydrogel precursor solution were generated in mineral oil, cured with a UV lamp, and sequentially brought into mechanical contact after passing a small aperture in the tubing that allowed the interstitial oil to escape. Right: Photograph of a fluidic artificial electric organ with 10 tetrameric gel cells generating 1.34 V. Scale bar = 1 cm. b, Open-circuit voltage and short-circuit current characteristics of fluidic artificial electric organ. Open-circuit voltages (red bars) scale linearly when tetrameric gel cells are added in series; short-circuit currents (blue bars) scale linearly when tetrameric gel cells are added in parallel. (Error bars show s.d., N = 3 except for 3×3, where N = 1). c, Plot of current and voltage in response to various external loads for one tetrameric gel cell (black squares), three cells in series (red circles), and three cells in parallel (blue triangles). d, Photographs of large complementary arrays of printed hydrogel lenses combining to form continuous series of 2,449 gels with serpentine geometry. Support gels are used for mechanical stability and do not contribute to the system electrically. Scale bar = 1 cm. e, Open circuit voltage and short circuit current characteristics of printed artificial electric organs as a function of the number of tetrameric gel cells in a series. f. Normalized current-voltage relations of various numbers of tetrameric cells added in series and in parallel. The voltage axis is normalized by the number of cells in a series; the current axis is normalized by the number of series that are arranged in parallel. All points fall on one curve, as expected for a scalable system.
The second feature that evolved in the electric organs of Electrophorus ensures simultaneous excitation of electrocytes along the entire ~1 m-long organ. Since nerve signals do not travel fast enough to simultaneously activate all electrocytes within the ~2 ms duration of a discharge, Electrophorus ensures synchronous signal delivery by slowing down the arrival of neural impulses to the parts of the organ closest to the command nucleus1.
The artificial electric organ also requires simultaneous activation across many gels to circumvent energy dissipation (Extended Data Fig. 2, Supplementary Information Section S2). Figure 2 shows two assembly strategies of an artificial electric organ, one based on fluidics and the other based on surface printing. The fluidic implementation automatically generated and positioned a series of gels sequentially using a programmable fluid dispenser (see Video 1). In this configuration, we prepared artificial electric organs with a maximum of 41 gels (Fig. 2a) and demonstrated that three gel columns in parallel delivered the expected triple current and power (Fig. 2b,c, Extended Data Table 2). Automated fluidic assembly thus makes it possible to fill devices in parallel and enables formation of a bundled artificial organ whose current and power scale with the number of gel columns analogous to the parallel electrocyte columns in the eel (Fig. 2b,c)1,2. While this first implementation of the fluidically-assembled artificial electric organ required three seconds per gel plug, state-of-the-art microfluidics systems can generate water-in-oil droplets of four different compositions15 at rates approaching 102 to 105 droplets per second16 and UV-induced polymerization of hydrogel particles in a flow-through reactor has been demonstrated at a rate of 25 gelled particles per second17. At these rates an artificial electric organ with 2,500 gels to generate 100 V may be assembled in less than 2 min. We demonstrated that by decoupling the fabrication from the assembly, for instance in the form of a tube pre-loaded with oil-separated hydrogel beads, assembly by pressure or external fields can be achieved within seconds (Video 1).
As an alternative to this fluidic assembly strategy, Figure 2d shows truly synchronous activation of an artificial electric organ by initiating mechanical contact between two complementary gel patterns over large arrays in a registered, parallel fashion. This design prints precursor solutions of ion-selective membranes in an array of lenses on one polyester substrate and precursor solutions of salt compartments on a second substrate in complementary patterns (Fig. 2d and Video 2). When overlaid after curing, the resulting hydrogel lenses instantly form a serpentine ionically conductive pathway (Fig. 2d) with a repeating motif of potentials that add to 110 V (Fig. 2e,f).
The third feature that evolved in the electric organs of Electrophorus enables maintenance and regeneration of large gradients of Na+ and K+ ions between the electrocyte cells and the extracellular space. ATP-dependent active transport by the Na+/K+-ATPase protein accomplishes this task by counteracting passive diffusion and rebuilding gradients after a discharge2.
In the artificial electric organ, we achieved maintenance of ionic gradients by physical separation between each of the hydrogels before actuation with the benefit that this design did not require energy expenditure. We regenerated the artificial organ after discharge by applying a current to the terminal electrodes in a manner similar to Gumuscu et al18 and recovered over 90% the original capacity over at least ten discharges (Supplementary Information Section S3, Extended Data Fig. 3).
In terms of performance as a power source, the electric organs of Electrophorus can generate 600 V by stacking parallel columns of 100 μm-thin electrocytes with membrane contact areas as large as 7,000 mm2 13, minimizing the resistance of a layer of electrocytes to ~ 0.1 Ω (Table 1) and enabling discharges of 100 W10. By comparison, the 110 V implementation of the artificial electric organ involved a conductive pathway along thousands of relatively thick hydrogel lenses with a small cross-sectional contact area (Fig. 2d), increasing the resistance to 115 kΩ per tetrameric gel cell and limiting the power output to 50 μW across 2,449 gels (Fig. 2e,f, Supplementary Information Section S4).
Table 1.
Comparison of parameters from the natural and artificial electric organ.
Source | Thickness of repeating unit* (m) | Cross-sectional area of electric organ (m2) | Open-circuit voltage per repeating unit (V) | Internal resistance of repeating unit (Ω m2) | Maximum power density generated by repeating unit (W m−2) |
---|---|---|---|---|---|
Live eels, anterior 6–10 cm of main organ, Nachmansohn et al. 1942 11† | (1.1 ± 0.2) × 10−4 | (2.3 ± 0.6) × 10−3 | 0.11 ± 0.01 | (6.8 ± 0.7) × 10−4 | 5.5 ± 1.4 |
Live eels, sections of main organ, Cox et al. 1946 13‡ | 10−4 | (3.4 ± 0.5) × 10−3 | 0.12 ± 0.01 | (5.1 ± 0.4) × 10−4 | 10.6± 2.0 |
Live eels, leaping, Catania 2016 8§ | 10−4 | 2.8 × 10−3 | 0.16 | 4.8 × 10−4 | 13.6 |
Gel cells, this work, 80° fold | 2.8 × 10−3 | 8.5 × 10−5 | 0.17 ± 0.01 | 0.27 ± 0.02 | 0.027 ± 0.002 |
Repeating unit refers to an electrocyte in the eel’s electric organ and to a tetrameric gel cell in this work.
Parameters are averages of the farthest anterior measurements for each of the 4 eels from that paper’s Table 2.
Cross-sectional area estimated from eels of corresponding length in Table 41 of Cox et al. 1946 13.
Parameters are averages of all measurements from that paper’s Table 1. Repeating unit thickness is an estimate reported in that paper. Open-circuit voltage values were calculated by multiplying reported electromotive force values by a factor of 0.77 (an average factor reported in the paper). Unit resistance and power density were estimated for each eel based on the approximation made in the paper that the external resistance was approximately equal to the internal resistance of the system.
Average of two experiments. Open-circuit voltage and short-circuit current extracted from highest position attained by each ascending eel where both parameters are shown in that paper’s Figure 4C. Cross-sectional area estimated from eels of corresponding length in Table 1 of Cox et al. 194613, repeating unit thickness estimated from anterior voltage distributions reported in Cox et al. 1940 30. Assumptions: eel tilt angle = 70°; electric organ begins 20% of reported eel length from its front tip11, resistance in the water tank from that paper’s Figure 3 is negligible.
To improve the performance, we devised flat hydrogel films on a patterned substrate and took advantage of a folding strategy developed for use in space to unfold solar panels (Fig. 3a, Extended Data Fig. 4, Supplementary Information Section S5). This Miura-ori fold makes it possible to stack a repeating series of thin films with large contact area in a single synchronized and self-registered motion19 (Video 3). Table 1 shows that stacking 0.7-mm thin films using an 80° Miura-ori fold yielded an artificial electric organ with a maximum power density of 27 ± 2 mW m−2 per tetrameric gel cell due to an approximately 40-fold reduction in resistance compared to lateral conduction through films of similar dimensions in the serpentine arrangement (Fig. 3b). As summarized in Table 1, the power density generated by this most efficient gel cell geometry is, however, still 2–3 orders of magnitude smaller than electrocyte layers in Electrophorus12,13.
Figure 3 |. Artificial electric organ morphologies based on thin hydrogel films.
a, Schematic and photographs of Miura-ori folding. A single motion compresses a two-dimensional array of panels into a self-registered folded state where all panels overlap, generating a one-dimensional sequence19. This morphology was used to generate flat and large contact areas between a series of thin gel films, which conducted ions from gel to gel through holes in the supporting polyester substrate. Scale bar = 1 cm. b, Area-normalized internal resistance (red) and maximum power (blue) per tetrameric gel cell with 0.7 mm thick gel films arranged either laterally in a manner that approximates the relative geometry of the serpentine implementation or in a Miura-ori-assembled stack (Error bars show s.e.m., N = 3) Section S4). The stack geometry imparts a 40-fold reduction in resistance and a corresponding 40-fold improvement in maximum power output. c, Flexible and transparent artificial electric organ prototype with the shape of a contact lens composed of a gel trilayer of high salinity gel (indicated by a false-colored section in red), anion-selective gel (yellow), and low salinity gel (blue) with a total thickness of 1.2 mm that produced an open-circuit voltage of 80 mV. Scale bars = 1 cm.
Bringing the design parameters of the artificial electric organ closer to the eel provides opportunities for further improving its performance. Electrocytes are thinner than the gel films shown in Fig. 3a by a factor of at least 7, and the absolute permeabilities of the ion-selective gel membranes used here were 10 times smaller than the ion permeabilities of electrocyte membranes (Extended Data Table 3)20. We showed that reducing the thickness of the hydrogel films by an order of magnitude increases the power density fivefold (Extended Data Fig. 5) and films of hydrogels as thin as a few hundred nanometers have been demonstrated in a different context21. Additionally, both the intracellular and extracellular compartments in the electric organs of Electrophorus are at physiological ionic strength of ~180 mM12 imparting low resistivity, while the artificial electric organ’s low-salinity compartment contains only 15 mM sodium chloride and contributes most of the system’s electrical resistance (Extended Data Fig. 5, Supplementary Information Section S6). In this regard, the work here highlights a non-obvious strategy the eel evolved to maximize power output. By developing electrocyte cells with two functionally different membranes such that, in the firing state, the posterior membrane is permeable specifically to sodium ions and the anterior membrane is permeable specifically to potassium ions (Fig. 1b), electrocytes convert the energy stored in opposing ion gradients of two different cations between physiological intracellular and extracellular solutions with low resistivity. In contrast, the broadly cation-or anion-selective hydrogel membranes in the artificial electric organ cannot distinguish between two different cations or anions, necessitating a gradient in overall ionic strength to establish a potential difference and hence inherently requiring one compartment with lower ionic strength and significantly higher resistivity than the other (see Supplementary Information Section S7). Therefore, development of ultrathin synthetic membranes with improved ion selectivity22 provides an opportunity to increase the power density of artificial electric organs.
Even before implementing these possible improvements, the unique material properties of artificial electric organs may inspire opportunities that cannot be imagined with conventional batteries. One example is the development of an electrically active contact lens made from hydrogel films. Contact lenses are commonly fabricated from hydrogels and prototypes with integrated displays and sensors have been reported23,24. To realize such functional lenses, an integrated and optically transparent power harvesting scheme would be enabling. Figure 3c illustrates a first step towards this vision. A single tri-layered lens generated a potential difference of 80 mV, hinting at the possibility of developing moldable, flexible, and optically transparent power sources for wearable and implantable devices5,25. Another example may be the use of artificial electric organs as deformable electric power sources in soft robots or other soft materials26.
The work presented here introduces the first implementation of an artificial electric organ from potentially biocompatible materials. By employing thousands of compartmentalized ion gradients as well as scalable fabrication combined with synchronized assembly of ion-selective membranes, these materials are able to generate total open-circuit potential differences in excess of 100 V and power densities of 27 mW m−2 per tetrameric gel cell. While these characteristics are only just approaching a useful power level for the lowest-power implant devices27, advanced designs – possibly with thinner and more selective membranes – combined with strategies to recharge and activate28 these artificial organs inside a living body29 are likely to increase their utility. Clearly, the eel’s electric organ demonstrates that organic electrical power sources inside living organisms can operate with exquisite power characteristics by utilizing metabolically available energy. If next generation designs can shorten the performance gap between the artificial and natural electric organ by one or two orders of magnitude then these artificial electric organs may open the door to metabolically sustained electrical energy for powering implants, wearables, and other mobile devices.
Methods:
Materials and Equipment:
We purchased all chemicals from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany) except for 40% acrylamide/N,N’-methylenebisacrylamide (henceforth “bis”) solution, which we purchased from Bio-Rad (Hercules, CA, USA) and P100 hydrophilic solution, which we purchased from Jonsman Innovation (Gørløse, Denmark). We purified all water to 18.2 MΩ cm with a PURELAB Flex II purifier (ELGA LabWater, Veolia, Paris, France). We performed all fluidic gel handling with a FIAlab-3500 multianalysis instrument (FIAlab Instruments Inc., Seattle, WA, USA) and an Infuse/Withdraw Pump 11 Elite programmable syringe pump (Harvard Apparatus, Holliston, MA, USA) in Tygon tubing (1/16” ID, 1/8” OD, McMaster-Carr, Elmhurst, IL, USA). We printed all patterns using a BioFactory 3D bioprinter (regenHU, Villaz-St-Pierre, Switzerland) with a custom print plate machined at the University of Fribourg. The transparent substrates were A4-sized uncoated polyester overhead transparencies with a thickness of 0.1 mm (Avery Zweckform #3555, Oberlaindern, Germany); we cut them using a Speedy 300 laser cutter (Trotec, Marchtrenk, Austria). We cured all gels using a Mineralight UV Display lamp (UVP, Analytik Jena, Jena, Germany) containing two 25-watt 302-nm or 365-nm UV tubes (Ushio Inc., Tokyo, Japan).
Characterization of artificial electric organs
The first and final gel compartment of each series of gels was a high-salinity reservoir gel. We inserted Ag/AgCl wire electrodes into the first and final gel of each series, or in identical high-salinity reservoir gels that we brought in contact with the terminal gels. We recorded voltages using a Tektronix DMM4040 digital multimeter set to high input impedance mode and short-circuit currents using a Keithley 2400 SourceMeter using a source voltage of zero (except where noted). We constructed I-V curves (Fig. 2c,f) by connecting a series of known load resistances to the batteries while monitoring the voltage across the load (Extended Data Fig. 6).
Fluidic artificial electric organ
The gel precursor solutions were all aqueous and had the following compositions: Low-salinity gel: 0.015 M sodium chloride, 0.045 M 2-hydroxy-4’-(2-hydroxyethoxy)-2-methylpropiophenone (henceforth “photoinitiator”), 5.5 M acrylamide, 0.067 M bis. High-salinity gel: 0.5 M sodium chloride, 0.045 M photoinitiator, 5.4 M acrylamide, 0.066 M bis. Cation-selective gel: 2.0 M 2-acrylamido-2-methylpropane sulfonic acid, 0.014 M photoinitiator, 3.7 M acrylamide, 0.045 M bis. Anion-selective gel: 2.0 M (3-acrylamidopropyl)trimethylammonium chloride, 2.75 M acrylamide, 0.034 M bis. We used McCormick food dye to differentiate the gels during experiments.
We connected the inlets of the FIAlab multi-position valve to sources containing the four different gel precursor solutions, and another containing mineral oil. A holding coil connected the main outlet of the valve to a syringe, which we used to withdraw liquids from the selected source into the holding coil. The main outlet was also connected to another source of mineral oil, which we pumped back through the outlet into a separate waste line to remove excess gel precursor. We programmed the FIAlab to withdraw a sequence of plugs of gel precursor solution ([high-salinity, cation-selective, low-salinity, anion-selective] x the number of repeats in the series), separating them with spacer plugs of mineral oil to prevent mixing of gels. After producing the desired number of plugs, we removed the collection tube and cured the gels under the UV lamp (365 nm) at a distance of 25 mm for 90 seconds. We connected the collection tube to a syringe filled with oil, then cut a notch-shaped opening on the opposite end of the tube, smaller than the thickness of the gel plugs. When we pushed the plugs past this notch, the interstitial oil escaped through the small opening and brought the entire sequence of gels into contact. We fabricated the first and final high-salinity gels separately and embedded the electrodes before curing. We connected these gels to the ends of the stacked gel sequence while applying light pressure and clamped them into place before electrical characterization. Video S1 illustrates the entire process.
Printed serpentine artificial electric organ for high voltages
The gel precursor solutions were aqueous and had the following compositions: Low-salinity gel: 0.015 M sodium chloride, 0.045 M photoinitiator 4.1 M acrylamide, 0.051 M bis, and 3.4 M glycerol. High-salinity gel: 2.5 M sodium chloride, 0.045 M photoinitiator, 5.1 M acrylamide, 0.062 M bis. Cation-selective gel: 2.0 M 3-sulfopropyl acrylate (potassium salt), 0.045 M photoinitiator, 1.9 M acrylamide, 0.055 M bis. Anion-selective gel: 2.0 M (3-acrylamidopropyl)trimethylammonium chloride, 2.75 M acrylamide, 0.034 M bis. Food dye from Städter was only used for photography and was absent during electrical recording experiments.
We printed two complementary arrays of 8 μL droplets of gel precursor solution on separate substrates in geometries that formed a single serpentine ionic pathway when overlaid (Fig. 2d). All the selective gel precursor droplets were printed onto the same substrate; the reservoir gel precursor droplets were printed onto the other one. After we finished printing onto a substrate, we immediately removed that substrate from the print plate and cured the gels under the UV lamp (302 nm) at a distance of 12 mm for 30 seconds. Once both substrates had been prepared, we overlaid them and applied a pressure of 16 kPa to the assembly to ensure contact between each of the gels in the series. In these experiments, voltages were collected with the SourceMeter using a source current of zero and currents were collected with a Keithley 6487 picoammeter.
80° Miura-ori folded artificial electric organ
The gel precursor solutions were the same as in the printed artificial electric organ. We laser-cut a 0.1 mm-thick polyester substrate with perforations in the Miura-ori pattern (80°, short edge of the parallelogram = 4 cm, long edge = 4.06 cm) and one 7 mm-diameter circular hole in each parallelogram positioned such that all the holes would overlap upon folding. We then attached PDMS pads with circular openings (depth = 0.3 mm, inner diameter = 10.5 mm) around each hole on both sides of the substrate, followed by manually pipetting the gel precursor solutions into each of the holes such that the PDMS reservoirs on both sides of the substrate were filled. We cured the solutions under a UV lamp (302 nm) at a distance of 12 mm for 30 seconds, then flipped the substrate over and cured the system for another 30 s. We peeled the PDMS reservoirs off the substrate before folding the substrates to bring the gels into contact (Fig. 3a). Painting a thin layer of deionized water onto both sides of each gel before folding prevented the gels from detaching from the substrate upon unfolding.
The power characteristics of the 80° Miura-ori folded geometry are compared with a lateral geometry of gels of the same dimensions as the ones shown in Figure 3a. To make the lateral analogues, we attached PDMS pads with circular openings to two planar substrates (depth = 0.7 mm, inner diameter = 10.5 mm) with the rings on each substrate spaced 5 mm apart from one another to approximate the relative geometry of the serpentine implementation. We filled the openings on one substrate with alternating membrane precursor solutions and the openings on the other with alternating reservoir precursor solutions, then cured them with the same parameters as the 80° Miura-ori gels and registered them as with the serpentine geometry.
Gel trilayer in the shape of a contact lens
The gel precursor solutions were the same as in the printed artificial electric organ. We made thin reservoirs by mounting substrates above a block of Teflon with double-sided tape, pipetted the precursor solutions into the reservoirs, and cured them under a UV lamp (302 nm) at a distance of 12 mm for 30 seconds to form thin (0.4 mm) sheets of the low-salinity, high-salinity, and anion-selective gels. We assembled the gel films into a trilayer and used a biopsy punch to obtain a circular cutout (diameter = 10.5 mm).
Extended Data
Figure 1.
Charged monomers used in charge-selective “membrane” gels. a, 3-sulfopropyl acrylate, a component of the cation-selective gel. b, (3-acrylamidopropyl)trimethylammonium, a component of the anion-selective gel.
Figure 2 |.
Self-discharge of artificial electric organ over time after contact between all gels with and without exposure to ambient air. Curves were fit with a single exponential decay function (dotted curves); the half-time for each was 40 min. The artificial electric organ was assembled as described in Supplementary Information section S3. Video 1 shows a fluidic implementation of the artificial electric organ which puts gels into contact sequentially rather than simultaneously. Large-scale implementations of similar sequential positioning schemes would be prone to power loss from gradient depletion.
Figure 3 |.
The artificial electric organ can be recharged. Experimental details in Supplementary Information section S3. a, Current versus time traces of ten discharges of a single tetrameric gel cell at short circuit following recharging. Initial discharge shown in black; subsequent discharges in the following order: red, blue, magenta, green, navy, purple, plum, wine, olive. b, Bar graph of normalized integrals of discharge curves.
Figure 4 |.
The printed 45° Miura-ori gel cell geometry. Dotted lines of a single color indicate gels forming a series. Different colors indicate parallel sequences. This fold geometry is scalable both in series for higher voltage output and in parallel for higher current.
Figure 5 |.
Internal resistance (black squares) and power density (red circles) of gel cells as a function of thickness of low-salinity gel. The thicknesses of all other gels were held constant at 1 mm.
Figure 6 |.
Equivalent circuit of an artificial electric organ connected to a load resistance. The elements within the dotted line represent the contribution of a single gel cell; these can be added in series or in parallel. The impedance of the voltmeter used exceeded 10 GΩ; current through this pathway was assumed to be negligible.
Table 1.
Selectivity of membranes considered in this work.
Reference | [Na+]in (M) | [Na+]out (M) | [K+]in (M) | [K+]out (M) | Membrane | Voc (mV) | PNa/PK* | PK/PNa* |
---|---|---|---|---|---|---|---|---|
Keynes and Martins- Ferreira, 195312† | 0.0131 | 0.172 | 0.1731 | 0.005 | Main organ, resting | −70.2 ± 2.9 | .0354 | 28.3 |
Main organ, peak | +55.1± 4.4 | 16.8 | 0.0597 | |||||
Sachs organ, resting | −78.1± 4.2 | .0182 | 55.1 | |||||
Sachs organ, peak | +61.9 ±2.1 | 31.3 | 0.0319 | |||||
Altamirano and Coates, 195732† | 0.0131 | 0.19 | 0.1731 | 0.005 | Sachs organ, resting | −84.0 ± 5.2 | .00789 | 127 |
Sachs organ, peak | +52.5 ± 2.6 | 11.7 | 0.0858 | |||||
Nakamura et al., 196533† | 0.0131 | 0.19 | 0.1731 | 0.005 | Sachs organ, resting | −84.9 ± 2.7 | .00366 | 140 |
Sachs organ, peak | +52.3 ± 4.1 | 13.6 | 0.0734 | |||||
Shenkel and Sigworth 199120 | 0 | 0.2 | 0.2 | 0 | Main organ, cell-attached patch, peak | +78.6 ± 6.7 | 21.3 | 0.0468 |
Main organ, inside-out patch, peak | +72.7 ± 11 | 16.9 | 0.0590 | |||||
Sachs organ, cell-attached, peak | +78.6 ± 6.6 | 21.3 | 0.0468 | |||||
Sachs organ, inside-out patch, peak | +73.6 ± 4.1 | 17.6 | 0.0568 |
[Na+]in (M) | [Na+]out (M) | [Cl−]in (M) | [Cl−]out (M) | Membrane | Voc (mV) | PNa/PCl | PCl/PNa | |
---|---|---|---|---|---|---|---|---|
This work | 0.015 | 2.5 | 0.015 | 2.5 | Anion-selective gel | −89.3 ± 0.6 | 0.0249 ± 0.0008 | 40.3 ±1.2 |
Cation- selective gel | +82.3 ± 1.5 | 29.5 ± 1.9 | 0.0343 ± 0.0024 |
Electrocyte membranes listed here are all the posterior, innervated membranes; anterior membranes maintain resting potential over the duration of a discharge12. Permeability ratios calculated using Equations S5 and S6.
When applicable, values here and throughout this work are presented as mean ± S.E.M.
Calculated from mean zero-current voltage.
Interior concentration values not provided; taken from Schoffeniels 195931.
Table 2.
Electrical characteristics from fluidic assembly of gel cells in series and parallel
Cell arrangement | Voc (mV) | Rint/L* (KΩ/mm) |
---|---|---|
One cell | 143 ± 3 | 4.8 ±0.8 |
Three cells in series | 438±5 | 14.6±0.9 |
Three cells in parallel | 143±3 | 2.0 ±0.1 |
Three series of three cells in parallel | 423 | 4.6 |
Internal resistance normalized by effective cell length, which accounts for differences in extent of compression. N = 3 for all values except final row.
Table 3.
Absolute permeability values of membranes considered in this work.
Reference | [Na+]in (M) | [Na+]out (M) | [K+]in (M) | [K+]out (M) | Membrane | Voc (mV) | PNa (m s −1) | PK (m s −1) |
---|---|---|---|---|---|---|---|---|
Shenkel and Sigworth 199120* | 0 | 0.2 | 0.2 | 0 | Sachs organ inside-out patch, peak | +75.3 ± 2.4 | (1.91 ±0.28) × 10−7 | (8.74 ± 1.10) × 10−9 |
[Na+]in (M) | [Na+]out (M) | [Cl−]in (M) | [Cl−]out | Membrane | Voc (mV) | PNa (m s −1) | PCl (m s −1) | |
---|---|---|---|---|---|---|---|---|
Anion- selective gel | −89.3 ± 0.6 | (6.71 ±4.10) × 10−10 | (2.61 ± 1.51) × 10−8 | |||||
This work | 0.015 | 2.5 | 0.015 | 2.5 | ||||
Cation- selective gel | 82.3 ± 1.5 | (2.03 ± 0.33) × 10−8 | (7.44 ± 1.71) × 10−10 |
Permeabilities calculated using Equations S3 and S8.
Values calculated from reported values of reversal potential (Er) and maximum current (Imax) from that paper’s Table 1.
Supplementary Material
Acknowledgements:
We are grateful to Prof. Barbara Rothen-Rutishauser and Prof. Alke Petri-Fink at the Adolphe Merkle Institute for the use of their BioFactory printer. Fritz Bircher's iPrint institute at the Haute École d’Ingénierie et d’Architecture Fribourg, particularly Florian Bourguet and Mathieu Soutrenon, graciously donated time towards adapting a printer for our use and helped us understand the intricacies of microvalve printing systems. Laser cutting was performed at Fablab Fribourg. Prof. Ullrich Steiner’s group, particularly Preston Sutton and Michael Fischer, provided instrumentation and advice related to impedance measurements. Research reported in this publication was supported by the Air Force Office of Scientific Research (grant no. FA9550–12-1–0435 to M.M., J.Y., D.S., and M.S.) and the National Institute of General Medical Sciences of the National Institutes of Health under award number T32GM008353, which funds the Cellular Biotechnology Training Program (T.B.H.S.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
The data that support the findings of this study are available upon request to the corresponding author.
Footnotes
Author Information:
The authors declare no competing financial interests.
References:
- 1.Bennett MVL Electric Organs in Fish Physiology (ed. Randall WSH. and D. J.) 5, 347–491 (Academic Press, 1971). [Google Scholar]
- 2.Gotter AL, Kaetzel MA & Dedman JR Electrophorus electricus as a model system for the study of membrane excitability. Comp. Biochem. Physiol. A. Mol. Integr. Physiol 119, 225–241 (1998). [DOI] [PubMed] [Google Scholar]
- 3.Xu J, Sigworth FJ & LaVan DA Synthetic Protocells to Mimic and Test Cell Function. Adv. Mater. 22, 120–127 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sun H, Fu X, Xie S, Jiang Y & Peng H Electrochemical Capacitors with High Output Voltages that Mimic Electric Eels. Adv. Mater. 28, 2070–2076 (2016). [DOI] [PubMed] [Google Scholar]
- 5.Kim D-H et al. Epidermal Electronics. Science 333, 838–843 (2011). [DOI] [PubMed] [Google Scholar]
- 6.Whitesides GM Assumptions: Taking Chemistry in New Directions. Angew. Chem. Int. Ed. 43, 3632–3641 (2004). [DOI] [PubMed] [Google Scholar]
- 7.Gallant JR et al. Genomic basis for the convergent evolution of electric organs. Science 344, 1522–1525 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Catania KC Leaping eels electrify threats, supporting Humboldt’s account of a battle with horses. Proc. Natl. Acad. Sci. 113, 6979–6984 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Xu J & Lavan DA Designing artificial cells to harness the biological ion concentration gradient. Nat. Nanotechnol. 3, 666–670 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Brown MV The electric discharge of the electric eel. Electr. Eng. 69, 145–147 (1950). [Google Scholar]
- 11.Nachmansohn D, Cox RT, Coates CW & Machado AL Action Potential and Enzyme Activity in the Electric Organ of Electrophorus Electricus (linnaeus): I. Choline Esterase and Respiration. J. Neurophysiol. 5, 499–515 (1942). [Google Scholar]
- 12.Keynes RD & Martins-Ferreira H Membrane potentials in the electroplates of the electric eel. J. Physiol. 119, 315–351 (1953). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Cox RT, Coates CW & Brown MV Electrical Characteristics of Electric Tissue. Ann. N. Y. Acad. Sci. 47, 487–500 (1946). [Google Scholar]
- 14.Pattle RE Production of Electric Power by mixing Fresh and Salt Water in the Hydroelectric Pile. Nature 174, 660–660 (1954). [Google Scholar]
- 15.Zeng S, Li B, Su X, Qin J & Lin B Microvalve-actuated precise control of individual droplets in microfluidic devices. Lab. Chip 9, 1340–1343 (2009). [DOI] [PubMed] [Google Scholar]
- 16.Bardin D et al. High-speed, clinical-scale microfluidic generation of stable phase-change droplets for gas embolotherapy. Lab. Chip 11, 3990–3998 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Young C, Rozario K, Serra C, Poole-Warren L & Martens P Poly(vinyl alcohol)-heparin biosynthetic microspheres produced by microfluidics and ultraviolet photopolymerisation. Biomicrofluidics 7, 044109 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Gumuscu B et al. Desalination by Electrodialysis Using a Stack of Patterned Ion-Selective Hydrogels on a Microfluidic Device. Adv. Funct. Mater. 26, 8685–8693 (2016). [Google Scholar]
- 19.Miura K Method of Packaging and Deployment of Large Membranes in Space. Inst. Space Astronaut. Sci. 1–9 (1985). [Google Scholar]
- 20.Shenkel S & Sigworth FJ Patch recordings from the electrocytes of Electrophorus electricus. Na currents and PNa/PK variability. J. Gen. Physiol. 97, 1013–1041 (1991). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ide T, Takeuchi Y & Yanagida T Development of an Experimental Apparatus for Simultaneous Observation of Optical and Electrical Signals from Single Ion Channels. Single Mol. 3, 33–42 (2002). [Google Scholar]
- 22.Feng J et al. Single-layer MoS2 nanopores as nanopower generators. Nature 536, 197–200 (2016). [DOI] [PubMed] [Google Scholar]
- 23.Lingley AR et al. A single-pixel wireless contact lens display. J. Micromechanics Microengineering 21, 125014 (2011). [Google Scholar]
- 24.Mansouri K, Medeiros FA, Tafreshi A & Weinreb RN Continuous 24-Hour Monitoring of Intraocular Pressure Patterns With a Contact Lens Sensor: Safety, Tolerability, and Reproducibility in Patients With Glaucoma. Arch. Ophthalmol. 130, 1534–1539 (2012). [DOI] [PubMed] [Google Scholar]
- 25.Lai Y-C et al. Electric Eel-Skin-Inspired Mechanically Durable and Super-Stretchable Nanogenerator for Deformable Power Source and Fully Autonomous Conformable Electronic-Skin Applications. Adv. Mater. 28, 10024–10032 (2016). [DOI] [PubMed] [Google Scholar]
- 26.Ilievski F, Mazzeo AD, Shepherd RF, Chen X & Whitesides GM Soft Robotics for Chemists. Angew. Chem. Int. Ed. 50, 1890–1895 (2011). [DOI] [PubMed] [Google Scholar]
- 27.Chandrakasan AP, Verma N & Daly DC Ultralow-Power Electronics for Biomedical Applications. Annu. Rev. Biomed. Eng. 10, 247–274 (2008). [DOI] [PubMed] [Google Scholar]
- 28.Enger CC & Simeone FA Biologically Energized Cardiac Pacemaker: in vivo Experience with Dogs. Nature 218, 180–181 (1968). [DOI] [PubMed] [Google Scholar]
- 29.Chin SY et al. Additive manufacturing of hydrogel-based materials for next-generation implantable medical devices. Sci. Robot. 2, eaah6451 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Cox RT, Rosenblith WA, Cutler JA, Mathews RS & Coates CW A comparison of some electrical and anatomical characteristics of the electric eel, Electrophorus electricus. Zoologica 25, 553–562 (1940). [Google Scholar]
Extended Data References:
- 31.Schoffeniels E Ion Movements Studied with Single Isolated Electroplax. Ann. N. Y. Acad. Sci. 81, 285–306 (1959). [DOI] [PubMed] [Google Scholar]
- 32.Altamirano M & Coates CW Effect of potassium on electroplax of electrophorus electricus. J. Cell. Comp. Physiol. 49, 69–101 (1957). [DOI] [PubMed] [Google Scholar]
- 33.Nakamura Y, Nakajima S & Grundfest H Analysis of spike electrogenesis and depolarizing K inactivation in electroplaques of Electrophorus electricus, L. J. Gen. Physiol. 49, 321–349 (1965). [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.