Implantable bioelectronic devices for photoelectrochemical and electrochemical modulation of cells and tissues

Abstract

Electroceuticals are bioelectronic devices that provide or modulate electrical or electrochemical signals to regulate physiological functions. In particular, devices designed for energy conversion are capable of transforming electrical energy into alternative forms of energy, such as heat or light, or vice versa, thereby enabling the photoelectrochemical and electrochemical modulation of biological systems, for example, to control muscle movement or cardiac rhythm. Such energy conversion approaches offer remote control and enhanced precision, surpassing the limitations of direct tissue and cell stimulation with traditional electroceutical devices, such as pacemakers, including mechanical mismatch at interfaces and wired communication. In this Review, we explore the fundamental principles of photoelectrochemical and electrochemical modulation of cells and tissues, emphasizing behaviour under physiological conditions. We then examine the development and application of implantable bioelectronics that use photoelectrochemical and electrochemical processes for modulation. Finally, we discuss future directions for energy conversion devices in implantable electroceuticals.

Introduction

Biological systems sustain themselves through complex energy conversion mechanisms, predominantly involving the transformation of external and internal energy into forms capable of driving cellular processes¹ (Fig. 1a). Electrochemical energy conversions within cells largely depend on the creation of ion gradients across cellular membranes. For example, in mitochondria, glucose is converted to adenosine triphosphate (ATP) in cellular respiration through a series of pathways — glycolysis, the citric acid cycle and oxidative phosphorylation² — all of which produce the electron-donating molecules nicotinamide adenine dinucleotide (NADH) and/or flavin adenine dinucleotide (FADH₂). Donated electrons then enter the electron transport chain, where a series of redox reactions occur, resulting in an electrochemical gradient that is used to produce ATP (chemical energy). On a larger scale, intercellular communications are established by neurons that convert chemical and electrical signals to transmit information to the brain. When neurons are stimulated, membrane depolarizations cause the propagation of action potentials (that is, rapid changes in voltage involving imbalances in ion gradients across the cell membrane) to the synaptic junction, inducing neurotransmitter release to pass the chemical signal to the next neuron³. However, in the sensory system, energy conversions are more complex as neurons convert external sensory cues to electrical energy through specialized transmembrane receptors. For example, in the visual system, light enters the retina and is received by photoreceptors, which convert optical signals into electrical signals in a process called phototransduction. Light is first sensed by photoreceptors in the retina, triggering the closure of sodium ion channels and leading to hyperpolarization; these signals are then transmitted to bipolar cells, which act as intermediaries, relaying the information from the photoreceptors to ganglion cells. Finally, ganglion cells generate action potentials to transmit signals to the visual cortex of the brain⁴. These pathways underscore the efficiency and complexity of biological energy transduction systems, which are crucial for all biological processes^1,5.

Fig. 1 |. Energy conversion in physical systems and biological entities and its implications for biomodulation device engineering.

a, Energy conversion processes in natural biological systems (left) include chemical-to-electrical energy conversion in mitochondria metabolism, electrical-to-chemical energy conversion in neuronal activation and optical-to-electrical energy conversion in retinal networks upon light stimulation, with implications for implantable biomedical devices (right). b, Cellular and tissue modulation may be direct, via ion channels and cellular receptors, or indirect, involving an energy transfer cascade prior to engagement with the cellular interface. These transformations include electrochemical transduction, chemical-to-electrical energy conversion and optical-to-electrical energy conversion. e^–, electron.

Leveraging this understanding, bioelectronic devices can be designed to interact with biological energy transduction systems (Table 1) directly or indirectly. These devices can modulate diverse biological signalling processes, from sensory responses to muscle movement and cardiac rhythm management^6,7. Devices interact directly with tissues by generating electrical or electrochemical outputs⁸. For example, endocardial pacemakers deliver electric impulses directly to the inner wall of the right ventricle through leads positioned in the endocardium. These electric pulses depolarize cardiac cells by modulating voltage-gated ion channels, initiating action potentials that induce coordinated heart contractions. By contrast, indirect interactions with tissues occur through devices that convert one form of physical energy into another before interacting with cells or tissues. For example, the non-genetic optoelectronic pacemaker features a silicon-based thin membrane capable of converting light into capacitive photoelectrochemical currents, which trigger heart contractions⁹. In both direct and indirect cases, bioelectronic devices interact with tissues, such as the skin, heart, gut and brain, facilitating physical-to-biological signal transduction through their engagement with various ion channels or receptors on the cell plasma membrane^10–13. For example, devices influence transmembrane proteins, such as voltage-gated or ligand-gated ion channels and G protein-coupled receptors, by electrical or chemical signals, light pulses, and mechanical forces as well as directly through the lipid bilayer to alter membrane biophysical characteristics, thereby enabling modulation of cellular activities and therapeutic interventions^14–17. This capability has led to the development of clinical tools, such as pacemakers, cochlear implants, spinal cord stimulators, insulin pumps and deep brain stimulators, with commercial success and positive impacts on health services¹⁸. However, these clinical devices are often bulky, such as those with large external batteries, which can increase the risk of infection, foreign body reactions and pocket haematomas¹⁹.

Table 1 |.

Energy conversion devices in bioengineering

Energy conversion	Examples	(Photo) cathode materials	(Photo) anode materials	Catalysts	Device operation principle	Biomedical applications	Ref.
Optical to electrical	Nanowire	n-type Si	p-type Si	Atomic Au	A p–i–n coaxial nanowire produces a photocathodic effect on the n-type Si shell to depolarize a neuron attached to it	Neuromodulation at the cellular level	78
	Nanowire	n-type Si	p-type Si	Atomic Au	A p–i–n coaxial nanowire produces a photocathodic effect on the n-type Si shell to depolarize neurons in mammalian spinal cord explants	Neuromodulation in mammalian spinal cord explants	79
	Inorganic membrane	n-type Si	p-type Si	Au nanoparticles	A p–i–n Si membrane produces a photocathodic effect on the n-type Si to depolarize neurons in forelimb primary motor cortex	Brain modulation for mouse forelimb movement	83
	Inorganic membrane	Illuminated nanoporous p-type Si region	Peripheral nanoporous p-type Si region and non-porous Si	NA	A nanoporous or non-porous Si membrane produces a photocathodic effect on the nanoporous Si to depolarize neurons in the sciatic nerve or cardiomyocytes in the heart	Mouse sciatic nerve modulation and rat cardiac stimulation	56
	Organic photovoltaic device	PC60BM	P3HT	NA	A P3HT and PCBM with TiN coating produces photo-capacitive stimulation in retinal ganglion cells	Photostimulation of retinal cells	76
	Organic semiconducting thin films	PTCDI	H2Pc	NA	An organic electrolytic photocapacitor based on PTCDI and H2Pc depolarizes sciatic nerve neurons	Chronic rat neuromodulation	30
	Inorganic membrane	Illuminated frontside p-type Si	Backside p-type Si	NA	Photocurrent caused by the built-in electric field in the Si solution heterojunction stimulates human bone marrow-derived mesenchymal stem cells	Mouse cranial defect repair	87
	Inorganic membrane	IUuminated nanoporous p-type Si region	Peripheral nanoporous p-type Si region and non-porous Si	NA	A nanoporous or non-porous Si membrane produces a photocathodic effect on the nanoporous Si to depolarize cardiomyocytes in the heart	Cardiac pacing in pigs	9
	Inorganic nanowires	Peripheral AuTiO2-x NW	AuTiO2-x NW	NA	AuTiO2-x NW arrays generate capacitive and faradaic photocurrent to stimulate cells in the retina	Vision restoration in mice and monkeys	82
	Phototransistors	NA	NA	Pt nanoclusters	A light-sensitive field-effect transistor integrated with eutectic GaIn alloy provides charge injections to retinal ganglion cells	Mouse vision restoration	53
	Inorganic NWs	TiO2	ZnTPyP	NA	A ZnTPyP self-assembled nanorod coated with TiO2 generates photo-capacitive current to induce action potentials in rat cortical neurons	Mouse neuromodulation	54
Chemical to electrical	Glucose fuel cells	Laser-induced graphene and Au nanoparticles	Laser-induced graphene and Au nanoparticles	Glucose oxidase (bio-anode) and laccase (bio-cathode)	A glucose oxidase-catalysed electrode oxidizes glucose to gluconic acid to generate an electric current	Mouse in vivo energy harvest and powering	106
	Glucose fuel cells	Pt/C catalyst on fibre electrode	CNT fibre electrode	Tetrathiafulvalene and glucose oxidase (bio-anode)	A glucose oxidase-catalysed electrode oxidizes glucose to gluconic acid to generate an electric current	Mouse in vivo energy harvest and powering	108
	Glucose fuel cells	Pt-CB/Nafion	CuO-MWCNTs-PEDOT:PSS	CuO (bio-anode)	A CuO-catalysed electrode oxidizes glucose to gluconic acid, providing electric power to the circuit for communication, recording and managing insulin release	Managing type 1 diabetes in mice	109
	Lactate fuel cells	Pt-Co nanoparticle-coated MDB-CNT	MDB-TTF-CNT/rGO/h-Ni	Lactate oxidase (bio-anode)	A lactate-catalysed electrode oxidizes lactate to pyruvate to generate an electric current	Human skin wireless sensing	22
	Mg–O2 bio-battery	Outer membrane-modified CNT/Pt	Mg	NA	Mg–O2 bio-batteries use O2 in the internal body fluid to produce an electric current for brain stimulation	Mouse brain stimulation	116
	Mg battery	Drug-loaded P(AM-co-SV) hydrogels	Mg	NA	A Mg battery enables reduction of viologen at the cathode, leading to expulsion of incorporated drugs for disease treatment	Mouse psoriasis treatment	27
	Mg battery	FeMn	Mg	NA	A Mg–FeMn battery offers sustained electrical stimulation for peripheral nerve regeneration	Rat neuroregenerative medicine	63
	Zn–O2 battery	CNT/Pt	Flexible Zn wire	NA	A tubular Zn–O2 battery provides the electric current for sciatic nerve stimulation	Regeneration of rat sciatic nerve	118
Electrical to chemical	NO generator	Pt-Fe3S4 Au-coated W microwire	W micro-wires electro-plated with Pt	Pt-Fe3S4	Nanoscale Fe3S4-based catalysts reduce nitrite or nitrate in biological systems, generating NO to trigger TRPV1 on HEK 293FT cells	Mouse neuromodulation	66
	CO generator	CoPc on oxygen-functionalized carbon paper	Pt	CoPc	The selective catalyst CoPc reduces CO2 dissolved in extracellular solution to CO, modulating cell signalling processes	Cellular modulation	122
	O2 generator	Stainless steel mesh with PtB	Stainless steel mesh with IrRuOx	PtB and IrRuOx	Electrolysis of water is used to produce O2 for a genetically engineered cell line that secretes therapeutic proteins	Provide O2 for xenotransplant cells in mice	124
	O2 generator	Pt	SIROF	SIROF	Electrochemical reaction	Provide O2 for cells	21
	H2 generator	H2-incorporated TiO2 nanorods	H2-incorporated TiO2 nanorods	NA	Sustainable photocatalytic reactions promote local glucose depletion and H2 generation	Treat wound healing in mice	127
	Electrogenetic interface	Pt wires	Pt wires	NA	Electrolysis of culture medium generates ROS, activating transgene expression in cells to produce insulin	Managing type 1 diabetes in mice	136

The conversion efficiency upper bounds are approximate values for cell-free or tissue-free devices (in which biointerfaces are not considered). CB, carbon black; CNT, carbon nanotube; CoPc, cobalt phthalocyanine; h-Ni, hierarchical Ni; H₂Pc, phthalocyanine; IrRuOx, iridium-ruthenium oxide; MDB, Meldola blue; MWCNTs, multi-walled carbon nanotubes; NA, not applicable; NW, nanowire; P3HT, poly(3-hexylthiophene); P(AM-co-SV), polyelectrolyte hydrogel, copolymerized from acrylamide and p-styrene-bipyridine monomers; PC60BM, [6,6]-phenyl-C61-butyric acid methyl ester; PEDOT:PSS, poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate); PtB, platinum black; Pt/C, platinum on carbon; PTCDI, N,N′-dimethyl perylenetetracarboxylic bisimide; rGO, reduced graphene oxide; ROS, reactive oxygen species; SIROF, sputtered iridium oxide; TTF, tetrathiafulvalene; ZnTPyP, zinc-tetra (4-pyridyl) porphyrin.

Innovations in energy harnessing and storage technologies have led to the development of devices that facilitate energy conversion, including high-efficiency fuel cells, electrolysis plants and solar cells^20–22 (Fig. 1a). Fuel cells convert chemical energy into electrical energy via redox reactions at the anode and cathode, and electrolysers use electrical energy to drive electrochemical reactions that produce valuable chemicals. Solar cells generate electric currents by creating and separating electron–hole pairs from light. These energy transduction principles can be harnessed to enable indirect modulation of biological activities and could benefit therapeutic electroceuticals^9,23–26 such as wearable fuel cells²⁷, in situ electrochemical pumps²⁸ and retinal prosthesis electrodes²⁹. This would require a cross-disciplinary approach integrating engineering and clinical perspectives.

Compared to direct modulation, indirect modulation offers several advantages as it enables remote control of cellular and tissue functions by converting primary energy forms into secondary ones, circumventing the limitations of direct modulation such as target proximity and biocompatibility issues (Fig. 1b). This approach leverages external energy conversion technologies to present a more versatile and less invasive option than direct interfacing by using a broad range of energy sources, including kinetic and metabolic energies⁵. For example, advancements in optical-to-electrical energy conversion processes allow the interfacing of deformable photocapacitor membranes to excite tissues without requiring bulky batteries, thereby enabling the non-invasive stimulation of cells through tissue-penetrating light^9,30. Additionally, hydrogel coating³¹, adhesive interfaces³² and stretchable electronics³³ can be integrated with indirect modulation devices to improve their biocompatibility and potentially provide minimally invasive, effective therapeutic interventions.

Energy conversion mechanisms pertinent to indirect bioelectronic modulation encompass various transitions, in particular, the conversion of electrical energy to chemical, thermal, mechanical, optical or magnetic energy within the device^34–39, and vice versa. These converted outputs primarily interact with cells and tissues through plasma membrane components. In this Review, we discuss how these principles can be leveraged to guide the design of indirect biological modulation tools. We focus on photoelectrochemical and electrochemical modulation of cells and tissues and outline strategies to develop bioelectronic devices that offer precise, minimally invasive therapeutic interventions⁴⁰.

Photoelectrochemical and electrochemical processes

Electrochemical processes involve the conversion of chemical energy into electrical energy (or vice versa) typically through redox reactions, where electrons move from an anode (where oxidation occurs) to a cathode (where reduction occurs), creating an electron imbalance that drives the flow of electrons through an external circuit. In some cases, such as capacitive processes, energy is stored electrostatically without redox reactions. By contrast, photoelectrochemical processes involve the conversion of optical energy to chemical or electrical energy. While electrochemical processes are powered by an external voltage or chemical reactions, photoelectrochemical processes rely on the internal electric field constructed when photons are absorbed by semiconductor materials.

Optical-to-electrical energy conversion principles

Photoelectrochemical modulation echoes photovoltaics, transforming light into electrochemical processes using semiconductors with saline interfaces, which in turn modulate cellular activity⁴¹. Both inorganic and organic semiconductors are used in photoelectrochemical systems. Inorganic semiconductors are typically silicon-based materials, valued for their high energy conversion efficiency and compatibility with complementary metal–oxide–semiconductor technology, which makes them ideal for integration into a wide range of electronic devices. Organic semiconductors, on the other hand, are often composed of organic pigment or conjugated polymers. The advantage of organic semiconductors lies in their flexibility and tunable bandgap. Photoelectrochemical systems⁴² use photodiodes or phototransistors to convert light into electrical energy, typically occurring at the interfaces of p-type and n-type semiconductors, or p–n junctions. The electrical energy powers subsequent electrochemical processes that are then delivered to target cells or tissues through attached microelectrodes. When semiconductors absorb light, excited electrons move from the valence band to the conduction band, leaving behind ‘holes’. A built-in electric field at the p–n junction drives apart these electron–hole pairs, leading to an imbalance of Fermi levels. When the semiconductor comes into contact with a saline solution, such as extracellular fluid, charge flows between the semiconductor and saline solution phases, generating an electric field and resulting in band bending at the semiconductor–saline interface^43,44. The built-in electric field and the electric field at the semiconductor–saline interface both influence the movement of charged ions and chemicals on the semiconductor surface, thereby affecting the operation of photoelectrochemical devices.

Photoelectrochemical devices can interface with cells and tissues under physiological conditions. Upon exposure to illumination from a light source, such as a light-emitting diode or laser, a capacitive current can be generated immediately within the electrochemical double layer of the photoelectrochemical device such as the optoelectronic cardiac pacemaker⁹, which induces the transient flow of ions in the extracellular fluid. This rapid change of ion concentration near the semiconductor alters the local electrochemical environment. Because cellular membranes are sensitive to ionic fluctuations, cells adjacent to the semiconductor can depolarize in response and initiate action potentials⁴⁵. Additionally, faradaic reactions on the semiconductor surface can influence redox processes at the semiconductor–biological interface through electron transfer reactions. Faradaic reactions occur when the light-generated voltage (photovoltage) exceeds the capacitive regime. These reactions can be further enhanced by the surface structure of the semiconductor or by the deposition of catalysts such as gold, which lower the activation energy for redox reactions and promote efficient charge transfer across the interface. The faradaic processes thereby alter the identity or concentration of redox species, such as reactive oxygen species (ROS), near cell or tissue surfaces^46,47. These alterations can be detected by cells and tissues through specific ion channels or membrane-bound receptors, which in turn influence cellular activation or inhibition by altering intracellular signalling pathways.

Various materials and device configurations use photoelectrochemical processes for biological stimulation. For example, organic photodiode interfaces employ organic semiconductor heterostructures as light-absorbing and transducing elements^48,49. Organic semiconductors use materials such as phthalocyanine (p-type) and N,N′-dimethyl perylenetetracarboxylic bisimide (n-type), which are efficient light absorbers with high absorbance coefficients of 10⁴ cm⁻¹ to 10⁶ cm⁻¹, enabling the creation of thin and flexible films suitable for bioelectronic interfaces^30,50. These semiconductors convert light to electrical energy to stimulate cells and tissues. Efficient electric stimulation is facilitated by creating an energy difference within the semiconductor, which leads to a light-driven separation of charges⁵¹. This energy difference is typically achieved through bulk or bilayer heterojunctions, which are domains or layers of different materials — one that donates electrons and another that accepts them. This creates a potential difference influencing the surrounding electrolyte such as extracellular fluid. The device itself may comprise organic semiconductors paired with single or dual conductors that function as the anode and cathode.

Phototransistors are distinct from conventional photodetectors that are based on photodiode or photoconductor setups⁵² (Fig. 2a). Unlike photodiodes or photoconductors, which primarily rely on the generation of a photocurrent directly from incident light, phototransistors combine the properties of a photodetector with amplification capabilities. When exposed to light, these transistors produce a photocurrent that boosts the drain current, enhancing charge delivery to cells or tissues via attached microelectrodes. For example, a phototransistor that integrates a high-resolution transistor array with 3D microelectrodes made from biocompatible liquid metal⁵³ can be interfaced with mouse retinal tissues to elicit action potential firings. This device offers a minimally invasive design owing to the low Young’s modulus (234 kPa) of the liquid metal electrodes, which reduces the risk of harming sensitive tissue layers.

Fig. 2 |. Optical-to-electrical, chemical-to-electrical and electrical-to-chemical energy transduction methodologies.

a–c, Conversion of optical to electrical energy is facilitated by the structure of photoelectrochemical semiconductors. Photodiodes comprised of diode junctions produce a photocathodic response (a flow of current or a voltage change) at cathode interfaces (part a). Catalyst integration on the photodiode membrane can augment photoelectrochemical performance. Phototransistors can be structured as either bipolar junction transistors (top) or field-effect transistors (bottom); however, they have not been used for biological modulation to date⁵². p–n diode junctions create staggered type II heterojunctions to segregate photoinduced charge carriers (part b). Porosity-based heterojunctions similarly induce staggered type II junctions for enhanced carrier separation⁵⁶ (part c). d–f, Conversion of chemical to electrical energy can be achieved with fuel cells, batteries and bio-batteries. The fuel cell mechanism involves open-system architecture to enable redox reactions, resulting in the generation of electric power⁵⁹ (part d). Conversely, the battery apparatus operates as a sealed system, accommodating the cyclical processes of charge accumulation and depletion⁶⁰ (part e). In biological analogues such as the electric eel, electrogenic cells (electrocytes) are arranged in series or parallel configurations creating biological voltaic arrays for current discharges⁵⁷ (part f). g–i, Administration of an electric potential to an electrode can result in a Faradaic reaction (part g) or invoke capacitive phenomena (part h), both of which are applicable to biomodulation. Addition of catalysts to these systems may potentiate the reaction by diminishing the requisite activation energy (part i). e^–, electron; E_C, conduction band edge; E_F, Fermi level; E_V, valence band edge; H⁺, proton; hν, photon energy; Ox, oxidant; Re, reductant. Parts b and c reprinted from ref. 56, Springer Nature Limited. Part f adapted from ref. 57, Springer Nature Limited.

The effectiveness of different semiconductors in bioelectronic devices is influenced by the bandgap structure of the semiconductors, their electrochemical properties and the surface characteristics of the materials. Strategies to enhance photoelectrochemical efficiency include the creation of porous structures to increase surface area and light absorption, surface engineering through catalyst deposition to enhance the kinetics of electrochemical reactions, and modification of bandgap properties to optimize light absorption and facilitate charge transfer^54,55. For example, a porosity-based heterojunction eliminates the need to control p-type or n-type doping across different material layers⁵⁶ (Fig. 2b,c). This involves introducing a nanoporous silicon layer next to a non-porous layer to achieve a staggered type II heterojunction, meaning that the conduction band and the valence band are misaligned in a way that the energy levels are staggered across the junction, facilitating the separation of charge carriers. This junction is akin to a typical p–n junction, resulting in similar photoelectrochemical properties. This approach allows the design of device configurations with high photostimulation efficiency.

Chemical-to-electrical energy conversion principles

Electrochemical energy conversion devices, such as fuel cells and batteries, transform chemical energy into electrical energy through redox reactions, enabling electric excitation of tissues as long as an electrode is connected, even if the energy conversion part or the power source is remote^57,58. The electrochemical dynamics at play in fuel cells and batteries are quantitatively described by the Nernst equation:

$$E=E^0-\frac{RT}{nF}\text{ln}\left(\frac{[\text{products}]}{[\text{reactants}]}\right)$$ (1)

where E° is the standard cell potential, R is the universal gas constant, T is the absolute temperature, n is the number of electrons transferred in the half-reaction and F is the Faraday constant (charge of one mole of electrons); [products] and [reactants] are the concentrations or pressures of the chemical species involved. The Nernst equation links the thermodynamic driving force of an electrochemical reaction to the resulting electric potential, establishing the equilibrium at which redox reactions occur. This potential is a critical metric for evaluating the capacity of the power source for bioelectronic stimulation. For instance, if the voltage is too low, it may not meet the threshold required to activate specific cellular or tissue responses, whereas higher voltages can successfully trigger the desired bioelectronic effects.

This equation is also pivotal in the design and application of battery-based or fuel-cell-based bioelectronic devices as it guides the optimization of ion transport and electrochemical processes for biological modulation^59,60. Understanding the spontaneity of redox reactions is essential for tuning these devices to achieve the desired bioelectronic stimulation. In fuel cells, energy conversion occurs when a fuel source is introduced to the anode, causing fuel molecules to oxidize, release electrons and generate ions⁵⁹ (Fig. 2d). The released electrons flow through an external circuit to the cathode, resulting in the production of electric power and, concurrently, the cathode is in contact with an oxidizing agent, such as atmospheric oxygen, that reduces and accepts electrons from the circuit. Fuel cells require a continuous consumption of fuel chemicals to generate electricity⁶⁰; by contrast, batteries store energy in a closed system by utilizing redox reactions and ion transport across electrodes (Fig. 2e), allowing them to provide self-contained power for bioelectronic devices. An example of a natural battery is the electric eel, which uses ion concentration gradients across its cell membranes to generate bioelectrical energy⁵⁷ (Fig. 2f). The mechanisms used by the electric eel have inspired the development of mouldable and transparent batteries that utilize ion concentration gradients across a hydrogel membrane, making them suitable for powering bioelectronic devices⁵⁷.

Bioelectronic therapeutic devices would greatly benefit from high-efficiency, biocompatible power sources. Conventional batteries are often bulky and contain potentially harmful materials and are thus challenging to integrate into physiological environments. Advanced batteries and fuel cells that feature stretchability, rechargeability, self-powering capabilities, biodegradability and high energy density^61–63 could seamlessly integrate within the body’s dynamic environments and provide a reliable and efficient energy supply for bioelectronic devices. This can be achieved through the development of novel materials, such as stretchable polymers, ionotronic hydrogels⁶⁴ and biodegradable composites, or by integrating other energy-harvesting technologies, such as piezoelectric and thermoelectric systems, to enable self-powering capabilities⁶⁵.

Electrical-to-chemical energy conversion principles

Many electrical-to-chemical energy conversions are facilitated by specific redox reactions such as denitrification reactions (nitrate to nitrogen reduction), which can be crucial for the precise modulation of biological tissues. For example, a Pt-Fe₃S₄ gold-coated tungsten microwire as the electrode implanted in the brain of mice can facilitate electrochemical reactions by transforming a current applied through the electrode (electrical energy) into chemical signals such as nitric oxide (chemical energy) (Fig. 2g,h). These chemical signals act with TRPV family member 1, a receptor of chemical stimuli in the central and peripheral nervous system, resulting in neuronal excitation in the mouse ventral tegmental area⁶⁶. Another example is using oxygen evolution reactions to generate oxygen on-site to support the viability of implanted cells in mice²¹. The efficacy of the electrochemical conversion relies on the reaction kinetics, particularly the relationship between the rate of an electrochemical reaction and the applied electric potential. The magnitude of current density traversing an electrode is characterized by the following expression:

$$j=j_0\left\{\text{exp}\left[\frac{{\alpha }_{a}zF}{RT}(E-E_{\text{eq}})\right]-\text{exp}\left[\frac{{\alpha }_{c}zF}{RT}(E-E_{\text{eq}})\right]\right\}$$ (2)

where j denotes the electrode current density, j₀ is the exchange current density, α is the transfer coefficient, z is the number of electrons involved in the electrode reaction, E is the electrode potential, E_eq is the equilibrium potential and n is the number of electrons involved in the reaction.

Electrocatalysts, catalysts that participate in electrochemical reactions, lower the activation energy required for a reaction to occur^67,68 (Fig. 2i). This speeds up the rate at which electrons can be transferred between the reactant, such as molecules in biological fluids, cell membranes or the extracellular matrix, and the electrode to provide efficient conversion between electrical and chemical energy. The quantitative relationship between the activation energy and the reaction rate is encapsulated in the Arrhenius equation:

$$k=A\text{exp}\left(-\frac{E_a}{RT}\right)$$ (3)

where k denotes reaction rate constant, A is a pre-exponential factor (related to the frequency of effective collisions), E_a is the activation energy, R is the universal gas constant and T is the temperature.

The ability of electrocatalysts to augment electrochemical signalling, such as using Pt-Fe₃S₄ to enhance denitrification reactions⁶⁶, highlights their enzyme-like behaviour, functioning as precise and efficient catalysts. This property offers significant potential for targeted and effective modulation of cells and tissues, particularly in complex systems such as neural and gastrointestinal networks. The performance metrics of an electrocatalyst are influenced by factors such as its surface area, electronic structure, and the magnitude of interaction with reactants⁶⁷. Manipulation of morphology and incorporation of heteroatom dopants, such as platinum or nitrogen, on the electrocatalyst surface are used for fine-tuning reactivity towards specific molecules^69,70. For example, a nanoporous structure can be incorporated to enhance the catalytic efficiency of iridium oxide in the oxygen evolution reaction by increasing the electrochemical surface area⁶⁹.

Electric fields can be used to indirectly activate chemical delivery. One approach is electrophoresis, in which an electric field is used to drive charged molecules over specific distances (~1–3 mm). This method can enhance drug delivery by increasing the permeability of biological barriers, such as skin or cellular membranes⁷¹. Similarly, iontophoresis is effective for transdermal drug delivery, involving the application of a low electric current to transport ions and charged drug molecules across the skin, enabling controlled release of molecules²⁷. Additionally, electrically responsive materials, such as conducting polymers and hydrogels, can be implanted in tissues and then indirectly activated upon electric stimulation to undergo volumetric changes and release encapsulated drugs precisely and on demand⁷². For example, applying a direct current to polypyrrole-doped membranes of drug-loaded polymers can reversibly alter their pore size, thereby controlling drug release rates from the membrane⁷³. These methods leverage the ability of the electric field to modulate ionic and molecular movement, providing a powerful tool for targeted and efficient chemical delivery in biomedical applications.

Implantable electroceuticals and photoelectroceuticals

Implantable electroceuticals are bioelectronic devices engineered to modulate tissue functions and treat a wide range of diseases by precisely regulating the body’s electrical signals through targeted electric stimulation. Recent innovations in energy conversion principles have opened new avenues for these devices, enabling wireless, lightweight and highly efficient operation through indirect methods. For instance, photoelectrochemical processes, such as those achieved with thin silicon membranes⁹, allow for the wireless and precise stimulation of cells and tissues; we refer to devices using this technology as ‘photoelectroceuticals’. These devices have broad potential applications, from treating chronic pain and regulating cardiac rhythms to modulating inflammation, offering a non-invasive and energy-efficient approach to controlling physiological functions.

Photoelectroceuticals

Electrode-based stimulation devices typically operate by directly delivering programmed electrical impulses to tissue. These devices often require wiring and leads and are limited by their invasiveness and bulkiness as well as the potential for mechanical failure and infection at the connection points⁷⁴. By contrast, many photoelectrochemical stimulation devices, such as nanoscale phototransducers and monolithic photodiode membranes, present wire-free, remotely controlled approaches that place minimal physical strain on the biological interface owing to their simple design⁷⁵. For example, photoelectrochemical devices can be implanted into target tissues and are activated with light stimulation to evoke either electrical or chemical signals that modulate biological processes. With precise control over the spot size and position of the light, photoelectroceuticals enable accurate spatial targeting and fine-tuning of the stimulation intensity and frequency. This versatility overcomes limitations and obstacles associated with pixelated electrode arrays such as manufacturing complexity and difficulties in portable wireless applications⁷⁶.

Nanoscale photoelectroceuticals.

Single-cell electrophysiology can be achieved with nanoscale phototransducers that convert light into localized electric currents, enabling precise targeting of individual cells for action potential generation. By directing light to specific regions, these phototransducers generate localized currents in the extracellular space or across the cell membrane, thus achieving targeted stimulation of the interfaced cells. For example, mesostructured silicon particles⁷⁷ and silicon nanowires⁷⁸ have been used to convert high temporal resolution optical input into electrical signals to generate action potentials and calcium influx in single cells (Fig. 3a). The surfaces of nanoscale phototransducers can be modified to enhance precise cellular targeting and improve their modulation efficiency, for example, in interfacing mammalian spinal cord explants⁷⁹. The silicon-based nanoscale photodiodes can be functionalized with antibodies against specific synaptic receptor components such as the GluA2 subunit of AMPA-type glutamate receptors. Through antibody–receptor interactions, the nanoscale photodiodes can be localized on the excitatory synapses of dorsal horn neurons and, when these photodiodes are remotely light activated, they evoke a significant increase in calcium event frequency and amplitude across the neuronal network. By modifying the nanoscale photodiodes with different antibodies, specific excitatory and inhibitory cell populations can be targeted to modulate sensory circuit processing. Furthermore, near-infrared light can be used to stimulate nanoscale photodiodes to enable deeper tissue penetration with low light scattering. Because of their ability to modulate neural circuits, nanoscale photodiodes have potential in the treatment of neurological conditions with dysfunctional signalling, such as chronic pain, by either enhancing pain signalling in the dorsal horn to address hyperalgesia or suppressing neural activity through activation of inhibitory interneurons. Leveraging drug-like delivery and precise cellular targeting, freestanding nanoscale transductors are useful for single-cell studies but may face challenges in large-area spatial control, electrical signal generation and heat dissipation, where planar photoelectroceutical devices offer greater robustness and integration.

Fig. 3 |. Application of photoelectrochemical constructs in biological systems modulation.

a, Coaxial silicon nanowires, with or without surface functionalization, are used to photoexcite neurons within the dorsal horn, activate the dorsal root ganglion, and train the cardiomyocyte or cardiac tissue to the targeted beating frequency⁷⁸. b, Porosity-based inorganic semiconductor membranes, manifesting as multi-layered heterojunctions, are used to optically pace the heart and photostimulate both sciatic nerves and cortical areas, achieving high spatial and temporal precision. Scanning transmission electron microscopy provides visualization of the coaxial nanowire structures and the nuanced components comprising the diode junctions of porosity-based heterojunctions⁵⁶. c, Multi-layered organic semiconductor apparatuses are implemented for the optical excitation of cells and tissues. This exemplary device uses a compact p–n bilayer film, consisting of phthalocyanine and N,N′-dimethyl perylenetetracarboxylic bisimide, overlaying a semi-opaque Au electrode. Upon irradiation with deep-red light, this assembly instigates formation of double layers, which, in turn, induce ionic currents culminating in biphasic pulses suitable for neural stimulation³⁰. d,e, Photoelectrochemical constructs have been applied for the creation of artificial retinas. These include 3D configurations of nanowire arrays, which involve Au-nanoparticle-decorated titania nanowires worked as photoreceptors⁸² (part d) and flexible and ultrathin phototransistor matrices integrated with soft liquid metal electrodes (part e). Incorporation of deformable materials, such as liquid metal, may reduce tissue invasiveness in these applications⁵³. e^–, electron; EGaIn, eutectic gallium-indium; H⁺, proton; hv, photon; Par-C, parylene C. Part a adapted from ref. 78, Springer Nature Limited. Part b adapted from ref. 56, Springer Nature Limited. Part c adapted from ref. 30, Springer Nature Limited. Part d adapted from ref. 82, Springer Nature Limited. Part e reprinted from ref. 53, Springer Nature Limited.

Planar photoelectroceuticals.

Micropatterned photoelectrode arrays represent earlier versions of planar photoelectroceutical devices, which can be integrated into degenerated retinas to replicate the role of natural photoreceptors^80,81. For example, such arrays can be implanted into the degenerated photoreceptor layer of mice retina⁸². A local photocurrent is then generated remotely to evoke electrical activity in the device-interfaced retina bipolar cells. This electrical signal results in downstream activation of retina ganglion cells to restore image-forming vision.

Thin-film flexible photoelectrodes (Fig. 3b) are adaptable to tissues, such as brain, retina, sciatic nerve and heart, and enable greater electrochemical current injection than nanoscale or micropatterned alternatives owing to their larger surface area^56,83. These films allow photoelectrochemical stimulation at optical intensities similar to those used in optogenetics (~1 mW/mm²)⁸⁴. They can be positioned and attached to tissues using van der Waals forces, capillary action or lock-in mechanisms, avoiding the need for external bioadhesives. Their high thermal conductivity, owing to reduced phonon scattering from material defects or grain boundaries, helps dissipate heat, reducing photothermal effects seen with nanoparticles⁸³. Silicon-based membranes are notable for their biocompatibility and potential bioresorbability^85,86, and engineered p–n (p-type/n-type), p–i–n (p-type/intrinsic/n-type) or porosity-based heterojunctions in their membranes yield substantial photocurrents and tissue-level stimulation. Upon remote activation by light, these membranes generate built-in electric fields to induce the localized ionic current that is transferred into the adjacent tissues to stimulate their activity. For example, a gold nanoparticle-modified p–i–n silicon membrane facilitates motor cortex stimulation⁸³, eliciting limb movement in anaesthetized mice; similarly a porosity-based heterojunction membrane can be applied to control heart rate in an isolated rat heart model using light pulses⁵⁶. These techniques show promise for neuromodulation and development of new treatments for neurological and cardiac disorders. Advances in organic device configurations include a wide-field spherical array for retinal stimulation that features thousands of photovoltaic pixels between a polymeric anode (poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate)) and a titanium and titanium nitride⁷⁶ cathode for high-resolution stimulation of retinal ganglion cells.

When implanting photoelectrode devices for long-term use, their in vivo stability and method of remote light activation should be considered. Silicon-based photoelectrodes are biodegradable and therefore suitable for short-term applications. Organic semiconductor devices (Fig. 3c), such as phthalocyanine–N,N′-dimethyl perylenetetracarboxylic bisimide devices, are stable in vivo, remaining functional for over 100 days when implanted on rat sciatic nerves, and therefore have potential for chronic neuromodulation³⁰. Deep-red light can penetrate up to 10 mm beneath the skin and is used for non-invasive activation of these devices, minimizing surgical risks and enhancing trans-tissue stimulation. In addition, photostimulation of an implanted device with 808 nm light promotes osteogenesis in cranial defects in a mouse model⁸⁷.

Monolithic devices enable multiscale and ‘random-access’ stimulation, where the area, location, timing and frequency are determined by the external illumination spots in a controlled yet flexible manner, offering spatiotemporal control over electrophysiological stimulation. These devices use porosity-based silicon heterojunctions and enable broad tissue coverage while providing precise modulation of individual cardiac cells⁹. Through activation with low light intensities, they can be applied to pace rodent heart tissue and for in vivo overdrive pacing in an adult pig heart. This technology allows high-resolution heart modulation at multiple sites, offering stimulation of 100 distinct positions within a 1 cm² area of a rat heart⁹. For clinical use, these devices can be integrated into catheter-like tubes for minimally invasive delivery through intercostal incisions⁸⁸, and paired with optical fibres for real-time, visually guided light delivery. This approach offers potential for heart modulation therapies and could transform the management of cardiac conditions. For example, this technology could be highly beneficial for patients undergoing cardiac resynchronization therapy who require multi-site pacing support. In cases where the initial lead placement for therapy is suboptimal, the random-access capability of a monolithic silicon-based pacemaker allows for precise, post-surgery readjustment of stimulation sites, reducing the risks associated with conventional pacemakers.

3D photoelectroceuticals.

While planar photoelectroceutical devices effectively interface with biological tissues, the transition to 3D device geometries offers the potential to significantly improve biointerface performance. This enhancement may be achieved through deeper tissue penetration, tighter extracellular junctions and, possibly even, intercellular integration, thereby facilitating more precise and efficient modulation of cellular environments. For example, 3D nanowire array-based structures are also being explored for translational photostimulation^82,89,90 (Fig. 3d). For example, gold nanoparticle-coated titania nanowire arrays can be implanted in the retina of mice and monkeys with induced photoreceptor degeneration to act as artificial photoreceptors⁸². The diseases being modelled in these studies include retinitis pigmentosa and age-related macular degeneration; both conditions result in the loss of photoreceptors, while other retinal cells and their projections to the brain remain functional. The device offers a spatial resolution of 77.5 μm and a temporal resolution of 3.92 Hz in ex vivo retinas. In blind mice, it enables the detection of gratings and flashing objects at light intensities as low as 15.70–18.09 μW mm^–2, with visual acuity of 0.3–0.4 cycles per degree. In monkeys, the arrays remain stable for up to 54 weeks and facilitate perception of a light beam with an intensity of 10 μW mm^–2 and a beam angle of 0.5° in visually guided saccade tasks. Therefore, the nanowire arrays can restore light perception and basic visual functions to blind participants and, furthermore, induce plasticity in the primary visual cortex in the form of long-term synaptic changes, suggesting the potential for the restoration of vision in individuals with photoreceptor degeneration. These findings suggest that, for patients with photoreceptor-degenerating diseases, such as age-related macular degeneration and retinitis pigmentosa, 3D nanowire array-based artificial photoreceptors may help restore certain levels of visual function.

In addition to diode-based devices, phototransistors are being explored in retinal prostheses^53,91 (Fig. 3e) using ultrathin photosensitive transistors integrated with flexible, 3D liquid metal electrodes⁵³ made of a eutectic gallium-indium alloy. Platinum nanoclusters on the electrode tips add nanometre-scale roughness and total surface area to the microelectrodes to enhance electrochemical properties and reduce impedance, improving charge injection into retinal ganglion cells⁵³. Conventional retinal prostheses use flat electrodes but, in patients with severe retinal degenerative diseases, the locally non-uniform retinal surfaces create an undesired gap between the retinal surface and the stimulation electrodes. The liquid state and 3D structure of the liquid metal electrode arrays offer a low Young’s modulus, reduced potential damage to retinal tissue and enhanced electrode proximity to target cells, which overcomes the limitations of flat electrodes. In vivo experiments in mice with retinal degeneration demonstrated the system’s potential for vision restoration through spatially precise stimulation⁵³. Additionally, unsupervised machine learning was used to analyse neural activity, confirming the device’s ability to generate precise neural responses to targeted light⁵³.

Challenges and solutions in photoelectroceuticals.

The biocompatibility and stability of photoelectroceutical materials, such as silicon and titania, can be improved for their extended usage. Although many of the devices are chemically inert or produce non-harmful byproducts⁹², the majority of inorganic materials have an elastic modulus in the gigapascal range, thus creating a mechanical mismatch with soft biological tissues, which have elastic moduli in the kilopascal range⁹³. Prolonged implantation of mechanically incompatible devices can result in damage to host tissues and ongoing inflammation⁹⁴. Additionally, the non-specific absorption of proteins on the device surface triggers immune cell responses and collagen deposition, promoting fibrosis at the device–tissue interface³². This fibrosis further impedes light penetration of the tissue and, therefore, the effective performance of the device. Several design alterations can be used to address the mechanical mismatch. For example, devices can be fabricated into ultrathin flexible membrane designs to reduce stress on curvilinear surfaces^9,95,96, porosification can reduce the apparent mechanical modulus⁵⁶, and a soft coating with conducting polymers or hydrogels can enable devices to interface seamlessly with the target tissues^97,98.

Chemical-to-electrical electroceuticals

Fuel cells.

The usage of chemical energy from human biofluids to power fuel cells was first proposed in the early 1960s⁹⁹. Human biofluids, such as sweat, lacrimal fluid and saliva, contain a variety of metabolites with high chemical energy that can be converted into electrical energy to power bioelectronics or bioelectrical modulation. Enzymatic fuel cells that utilize biological enzymes as catalysts to drive redox reactions operate in a similar manner to conventional fuel cells that use metallic electrocatalysts. For example, fuel cells equipped with glucose oxidase allow consumption of body glucose to produce electrical energy^100,101. At the anode, glucose oxidase enzymes oxidize glucose into gluconate to generate protons. Additionally, an electrode is located at the cathode, where oxygen accepts these protons and generates water, resulting in a net electric current¹⁰². Materials such as carbon nanotubes, graphene composites or conductive nanoparticles are used to facilitate efficient electron transfer between the enzyme redox centre and the electrode^103–105. For example, adding gold nanoparticles to laser-induced graphene electrodes accelerates electron transfer, resulting in higher electric outputs with an open-circuit potential of 0.77 V (ref. 106). When combined with biodegradable poly(lactic-co-glycolic acid), this biofuel cell can be made biocompatible and can serve as a power source for low-power implantable electronic devices. In addition to electrode modification, ceramic materials can be used (such as cerium dioxide) as electrolytes in glucose fuel cells to improve their adaptability for implantation purposes by enabling miniaturization of the devices and thermal sterilization¹⁰⁷. However, implanting enzymatic biofuel cells in tissues poses challenges related to long-term operational stability under physiological conditions due to biofouling, degradation of enzymes and loss of catalytic activities. Electrodes on implanted biofuel cells can be coated with hydrophilic zwitterionic polymers to prevent biofouling¹⁰⁸. Replacing glucose oxidase with electroactive metal oxides, such as cupric oxide, also offers a way to improve stability. By using cupric oxide, multi-walled carbon nanotubes and poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) on the anode side, the fuel cell operates efficiently within the 4–50 mM physiological blood glucose range¹⁰⁹ (Fig. 4a). Additionally, non-enzymatic biofuel cells can also achieve power densities of 0.45 mW cm^–2 and an open-circuit voltage of 0.6 V with 10 mM glucose. Non-enzymatic biofuel cells function by using nanomaterials such as metal oxides as artificial enzymes to catalyse electrochemical reactions rather than relying on biological enzymes for the conversion of biofuel into electrical energy. In vitro experiments using cells genetically modified to release insulin in response to light were used to test the efficacy of non-enzymatic biofuel cells in closed-loop glucose modulation. Under physiological conditions, the electric field generated by the device is sufficient to produce blue light, which optogenetically depolarizes the cell membrane, triggering local insulin release¹⁰⁹.

Fig. 4 |. Application of chemical-to-electrical energy transduction systems in biological systems modulation.

a, A non-enzymatic implantable, glucose-responsive fuel cell monitors blood glucose levels and can be remotely activated to deliver optostimulation or electrostimulation to cells, triggering insulin release. The fuel cell catalyses blood glucose to gluconate at the anode, generating electrons (e^–) for the power circuit and protons (H⁺) for the cathode. The produced electrical energy is then wirelessly transmitted to external monitoring devices that regulate the fuel cell’s operation according to cellular glucose concentrations¹⁰⁹. b, A self-sufficient electronic skin extracts energy from the human body to facilitate data communication to a mobile device through Bluetooth for personalized metabolic monitoring, including for urea and glucose²². c, A wearable iontophoretic patch is powered by a Mg battery for transdermal dexamethasone (Dex) administration²⁷. d, A compact, self-powered, biodegradable device is tailored for nerve restoration in neuroregenerative therapies. This device comprises a Mg–FeMn galvanic cell, which supports axonal outgrowth and calcium activity of dorsal root ganglion neurons, and porous polycaprolactone (PCL) and a copolymer of poly(l-lactic acid) and poly(trimethylene carbonate) (PLLA-PTMC) that provide a soft and biocompatible interface⁶³. e, Schematic of an implantable, tubular Zn–air (Zn–O₂) battery configured to wrap around a nerve and act as an in vivo power supplier to promote regeneration of the injured sciatic nerve¹¹⁸. CNT, carbon nanotube; PLGA, poly(lactic-co-glycolic acid); PVA/PBS, polyvinyl alcohol/phosphate-buffered saline. Part a adapted from ref. 109, CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Part c reprinted from ref. 27, Springer Nature Limited. Part e adapted from ref. 118, Wiley.

In addition to glucose, lactate in sweat has been explored as a promising alternative energy source for fuel cells and has primarily been used for real-time biological sensing^22,110. Lactate is found in high concentrations in sweat and is easily oxidized by lactate oxidase to pyruvate¹¹¹. For example, a battery-free, fully perspiration-powered electronic skin equipped with lactate biofuel cells enables continuous monitoring of metabolic biomarkers such as urea, NH₄⁺ and pH²² (Fig. 4b). By modifying the electrode with zero-dimensional to 3D nanomaterials, the lactate biofuel cells achieve a power density of 3.5 mW cm^–2 and show stable performance throughout 60-h operations in non-invasive human metabolic monitoring. Direct application of fuel cells for therapeutic modulation is also feasible. For example, flexible fuel cells can be constructed to convert chemical energy from bodily or externally supplied fructose into low-intensity direct currents and used to accelerate wound healing¹¹². Moreover, biofuel cells using fructose as energy input can aid in transdermal drug delivery via iontophoresis¹¹³. These cells create an ionic current across the skin from the bio-anode to the bio-cathode, enabling the controlled release of ascorbyl glucoside from a hydrogel into the skin.

Fuel cells rely primarily on a sufficient supply of biofluids, which limits their ability to sustain stable and continuous operation⁵⁸. Alternatively, battery systems with high operating voltage and stable power output can provide energy sources for bioelectronics. Unlike fuel cells, batteries can convert already stored chemical energy into electrical energy by using the surrounding biofluids as electrolytes. However, conventional lithium-ion batteries are not ideal for biomedical applications as they often contain toxic materials such as heavy metals and inherently flammable electrolytes¹¹⁴. Electrode materials for implantable batteries should be biocompatible and biodegradable upon contact with body fluids. Magnesium is a vital mineral involved in more than 300 metabolic processes and is thus a promising candidate for electrode materials¹¹⁵.

Magnesium-based batteries.

Magnesium-based batteries, known for their lightweight and biocompatibility, have demonstrated potential in applications including neuromodulation and iontophoresis^63,116. In a typical magnesium battery system, magnesium is oxidized to form magnesium hydroxide at the anode and oxygen from biofluids is reduced to hydroxide at the cathode. The cathode can include polyelectrolyte hydrogels to provide a stable and redox-controlled response to electro-stimuli²⁷. In a typical iontophoresis device, the reduction of viologen (a reversible redox-active mediator) in response to electrical energy triggers drug release²⁷ (Fig. 4c). The generated energy density can reach 3.57 mWh cm^–2, which is higher than that of typical self-powered iontophoresis devices. Furthermore, a biofluid-activated battery that features magnesium as the anode and an iron-manganese alloy as the cathode within a porous polycaprolactone scaffold can generate an electric current along the conduit. The generated electric field accelerates local neuroregeneration and promotes functional recovery of sciatic nerve tissue in a rat model⁶³ (Fig. 4d). However, magnesium-based electrodes typically exhibit a high corrosion rate, on the order of hours, in vivo¹¹⁷, leading to the dissolution of magnesium and the production of hydrogen, which can cause gas poisoning and blinding of the magnesium anode.

Zinc-based batteries.

Biocompatible zinc-based batteries with a tubular shape exhibit good biosafety, higher power density and lower corrosion rate (days to weeks) than magnesium batteries, facilitating electrical stimulation of nerve structures in sciatic nerves in rats¹¹⁸. During in vivo discharge, the zinc anode undergoes oxidation to Zn²⁺ and oxygen is reduced to hydroxide ions at the carbon nanotube/platinum cathode. The porous structure of the carbon nanotube at the cathode also enhances the exchange of oxygen, nutrients and waste. The zinc–oxygen battery achieves a specific volumetric energy density of 231.4 mWh cm^–3 and has been used to promote regeneration of the injured sciatic nerve in rats (Fig. 4e). Biocompatible zinc-based batteries can also be incorporated into stretchable substrates or integrated with textiles and garments to power bioelectronic devices⁶¹. The application of chemical to electrical energy conversion in bioelectronics therapy is still in its infancy. Addressing challenges, such as the leaching of materials, biofouling and interference from biological species, is crucial for the translation of these technologies.

Biomimetic batteries.

The electric eel can act as a model for a biomimetic soft material-based battery system. The eel’s natural electricity-generating mechanism can be mimicked by the stacking of hydrogel droplets with ion gradients. This system generates electrical energy through the movement of cations, including Na⁺ and K⁺, and anions across selectively permeable membranes, producing an open-circuit voltage of 127 mV and a short-circuit current peaking at 2.2 μA. A miniaturized soft power source can be produced by shrinking the volume of a power unit by over 10⁵-fold compared to other eel-inspired designs⁵⁷ and achieving a power density of approximately 1,300 W m^–3 (ref. 119). These biomimetic power sources generate ionic currents that stimulate intracellular calcium waves, enabling the modulation of neuronal networks. Their biocompatibility and functionality make them suitable for integration into living organisms for applications such as powering bioelectronic devices, enabling neural stimulation for therapeutic purposes, facilitating tissue regeneration, or powering implantable sensors for continuous health monitoring and disease treatment.

Electrical-to-chemical electroceuticals

Chemicals allow finely tuned, long-lasting modulation of cellular systems; however, the precision of direct chemical delivery is often limited¹²⁰. The delivery of messenger molecules frequently leads to off-target effects and provides limited control over release kinetics¹²¹. Precise treatment of diseases depends on the real-time release of target molecules at specific positions. To meet this need, the electrocatalytic route enables the generation of chemical messengers with tunable release kinetics⁶⁶. Electrical-to-chemical electroceuticals electrocatalytically generate bioactive chemicals to modulate tissues. For example, in electroceutical devices, an electric field can be applied to iron sulfide-based catalysts at the cathode to reduce externally infused nitrogen dioxide (NO₂⁻) into bioactive nitric oxide⁶⁶ (Fig. 5a). Catalytic activity and selectivity toward nitric oxide are further enhanced by adding platinum, a widely studied NO₂^– reduction catalyst, to the iron sulfide catalyst surface. In addition to nitric oxide, carbon monoxide enables localized cell signalling activation through an on-demand electrochemical carbon dioxide reduction reaction¹²². The ability to fine-tune carbon monoxide release kinetics facilitates systematic exploration of neuronal signalling processes such as non-adrenergic non-cholinergic neurotransmission influenced by this molecule¹²³.

Fig. 5 |. Applications of electrical-to-chemical energy transduction in biological systems modulation.

a, Electrochemical systems facilitate the conversion of NO₂^– to NO using Fe₃S₄ and Pt-Fe₃S₄ nanocatalysts. NO gas is produced through the reduction of NO₂ on the nanocatalysis surface (not shown in the diagram), while water oxidation at the anode produces O₂ (ref. 66). b, On-site electrocatalytic oxygen generation (ecO₂) involves an oxygen evolution reaction on a nanostructured sputtered iridium oxide film to sustain cell viability and maintain therapeutic secretion at a high cell density¹²⁴. c, Electric current can be delivered to induce transgene expression in modified mammalian cells. Cells modified for inducible expression of the NRF2 transgene are cultured on a 24-well plate and electrically stimulated using platinum electrodes. This triggers gas formation and NRF2-mediated and KEAP1-mediated antioxidant reactions to deliver NRF2 into cell nuclei, resulting in NRF2-mediated gene expression. In the absence of electrostimulation, NRF2 is degraded by the proteasome. d, This electrogenetic framework has been used for the treatment of type 1 diabetes in mice, as the engineered cells can be implanted in the dorsal area of mice and then electrically stimulated by acupuncture needles with 4.5 V voltage¹³⁶. AA, 1.5 V AA battery; DART, direct current-actuated regulation technology; e^–, electron; GOI, gene of interest; H⁺, proton; pA, polyadenylation signal; ROS, reactive oxygen species. Part a adapted from ref. 66, Springer Nature Limited. Part b adapted from ref. 21, Springer Nature Limited. Parts c and d adapted from ref. 136, Springer Nature Limited.

Electrocatalytic water splitting.

In vivo electrocatalytic processes have the potential to control and modulate biological functions, such as tissue oxygenation and metabolic processes, offering new approaches to therapeutic applications. Electrocatalytic water splitting using electrolytic cells results in oxygen and hydrogen products. Because these devices can generate oxygen on-site, they can be used to alleviate hypoxic conditions in transplanted tissues or cells to enhance their survival and function¹²⁴. This is particularly relevant for tissue engineering and regenerative medicine because sufficient oxygenation is crucial for the viability of engineered tissues and organs. Additionally, an electrocatalyst can be used¹²⁵ to avoid the need for high overpotentials in neutral environments and the risk of generating toxic byproducts such as chlorine¹²⁶. For example, a wireless, battery-free device can generate oxygen and provide immune protection for therapeutic xenotransplants in vivo¹²⁴ (Fig. 5b). Key components of the device include a membrane electrolyser assembly for water vapour electrolysis, which generates oxygen and hydrogen through an ion exchange membrane layer sandwiched between anode and cathode catalyst membranes, a gas diffusion chamber for oxygen storage, and an immune-isolating cell-housing chamber that enables oxygen permeation and protects the encapsulated cells from the host’s immune response. The device uses iridium-ruthenium oxide as the anodic catalyst and platinum black as the cathodic catalyst within the membrane electrolyser assembly to facilitate water splitting into oxygen and hydrogen. The generated oxygen is then stored in a silicone-based gas diffusion chamber and supplied to the encapsulated cells through a polydimethylsiloxane polymer membrane, ensuring the cells have adequate oxygen to thrive. Similarly, an electrocatalytic on-site oxygenator can support implanted therapeutic cells by providing controlled oxygen generation directly at the implantation site²¹. The oxygenator platform uses nanostructured sputtered iridium oxide as an anode catalyst for the oxygen evolution reaction at neutral pH. This choice of material ensures biocompatibility as lower oxygenation onset and selective oxygen production are possible without generating toxic byproducts. The electrocatalytic on-site oxygenator can also sustain high cell loadings (over 60,000 cells per mm³) in hypoxic conditions for implantation both in vitro and in vivo. Furthermore, in animal models, the device supports cell viability and function without eliciting adverse immune reactions, highlighting its potential use in cell-based therapeutic interventions.

Hydrogen, the other product of water splitting, can also be used for biomedical treatment. For example, titanium oxide nanorods can serve as photocatalysts for hydrogen generation. Through the oxidation of glucose as a sacrificial agent, the generated hydrogen from the reduction half-reaction promotes cell proliferation and migration in a diabetic wound model in mice, providing a safe and efficient treatment approach for diabetic foot ulcers¹²⁷.

Metabolic pathway modulation.

In vivo electrocatalysis can be harnessed to influence metabolic pathways directly, through implanted bioelectronics, by providing essential substrates or removing metabolic byproducts¹²⁸. This approach could be instrumental in managing metabolic disorders such as diabetes, in which precise control over glucose metabolism and insulin sensitivity is required. Additionally, in electrodynamic therapy, various nanoparticles can be engineered to generate ROS independently of the microenvironment¹²⁹ when activated by an electric field^130,131. Bioengineering techniques, including conjugation of nanomaterials with extracellular vesicles, can amplify therapeutic effects by prolonging blood circulation and accumulation at infection sites, promoting systemic anti-tumour immunity and combating infections^132,133. Moreover, the electrochemically generated ROS can serve as a bridge between electronics and genetically modified living systems^134,135. In a proof-of-concept experiment for type 1 diabetes, acupuncture needles were used to generate ROS by applying a 4.5 V direct current for 10 s. The electrogenetically engineered cells implanted in diabetic mice successfully increased insulin production and modulated blood glucose levels in response to the generated ROS¹³⁶ (Fig. 5c,d).

Drug delivery.

The electrochemical mechanisms in bioelectronic devices can also be designed for drug delivery. For example, drugs can be loaded into a valve in the device, then a wireless electrical energy input is used to trigger the corrosion of valve structures to release the drug. In the rat model, these wirelessly programmable devices can be implanted under the skin to deliver local anaesthetics for pain mitigation¹³⁷. Biological modulation using electric field-enabled drug delivery leverages techniques such as electrophoresis and iontophoresis to precisely control the administration of neuroactive substances. For example, organic electronic ion pumps can deliver Ca²⁺ ions to neuronal cells²⁸. By applying an electric field to activate ion pumps, these devices can induce localized Ca²⁺ signalling, modulating activity in neuronal cells and offering potential treatments for neurological disorders such as epilepsy. Electrophoretic delivery systems can also be used to administer neurotransmitter γ-aminobutyric acid for seizure control¹³⁸. This method provides precise temporal and spatial drug delivery, enabling effective management of epileptic seizures with minimal systemic side effects. The electrophoretic methods have already been used to deliver neurotransmitters to modulate sensory functions in mammalian models¹³⁹. This approach involves the precise control of electrophoretic migration to release glutamate for the selective stimulation of targeted cells in the cochlea, demonstrating the potential to restore sensory functions. These examples illustrate how electric fields can be used to enhance the precision and efficacy of neuromodulation therapies, opening new avenues for the treatment of various neurological conditions with targeted drug delivery systems.

Outlook

Precise intracellular control

Bioelectronic modulation techniques often struggle with precise intracellular control as most tools operate extracellularly¹⁴⁰. The introduction of electrocatalytic processes into devices may provide direct and controlled intracellular signalling at the molecular or organelle level. For example, electrocatalytic devices may modulate mitochondrial activity by directly influencing oxidative phosphorylation through the modulation of the redox states of key components such as cytochromes² (Fig. 6a). Additionally, targeting the electron transport chain by using intracellular devices to influence the activity of complex I or complex IV could enable regulation of cellular respiration and oxidative stress. Diseases such as diabetes and metabolic syndrome are linked to impaired mitochondrial function. By regulating respiration at the level of complex I or complex IV, intracellular devices could help restore normal energy metabolism and reduce oxidative stress, potentially alleviating the symptoms of these conditions.

Fig. 6 |. Improving control and integration of bioelectronic devices.

a, The precision of intracellular modulation might be improved at the organelle level, in which genome engineering has a crucial role in the control of cellular signalling². b, For example, the eCRISPR (electrogenetic clustered regularly interspaced short palindromic repeats) system uses targeted electrochemical gradients to direct gene editing, thus enabling acute activation of gene responses at the bioelectronic interface¹⁴¹. c–e, Progress towards less invasive implantation techniques includes delivery of hydrogel electrodes via a catheter into coronary veins for in situ formation and pacemaker integration¹⁴⁴ (part c), catheter-like devices to deploy and attach silicon-based cardiac stimulators to pig hearts enabling minimally invasive photostimulation⁹ (part d), and distinguishing between non-targeted cells and cells targeted for gene delivery expressing membrane-bound horseradish peroxidase and a fluorescent protein, resulting in polymerization at the cell surface¹⁴⁶ (part e). f, Future development of living bioelectronic systems can combine bioelectrical, biomechanical and biogenic materials to create multifunctional devices⁹⁸. Hydrogels offer tissue-like biomechanical properties to reduce the mechanical mismatch with tissue and increase the fidelity in signal recording⁹⁷. Functionalization with biogenic compounds, such as bacteria and cells, offers the capability for tissue immunomodulation and therapeutic intervention. Part b adapted from ref. 141, Springer Nature Limited. Part c adapted from ref. 144, Springer Nature Limited. Part d adapted from ref. 9, Springer Nature Limited. Part f adapted with permission from ref. 98, American Association for the Advancement of Science.

An electrogenetic clustered regularly interspaced short palindromic repeats (CRISPR) system has been developed in the bacterial models Escherichia coli and Salmonella enterica¹⁴¹, which can transmit electronic information into biological systems through redox mediators, enabling the precise control of CRISPR-based information such as producing specific proteins, toxins or metabolites (Fig. 6b). This method amplifies signals and filters noise using controlled redox reactions and CRISPR systems, providing a new way to interpret and manipulate biological functions electronically, with implications for biohybrid microelectronic devices and the broader fields of synthetic biology and bioelectronics. These emerging strategies in electrocatalytic bioelectronics hold promise for precise intracellular signalling control.

Seamless integration and minimally invasive device deployment

The seamless integration and minimally invasive deployment of in vivo implantable devices for biological modulation^142,143 have particularly been enabled by injectable hydrogel electrodes, genetically targeted chemical assemblies and conducting polymers. For example, in a porcine model, injectable hydrogel electrodes have shown promise for the treatment of ventricular arrhythmias¹⁴⁴, a leading cause of sudden cardiac death (Fig. 6c). This approach involves inserting hydrogel electrodes into coronary veins; the flexibility of hydrogels enables electrode penetration of mid-myocardium regions inaccessible to traditional pacing methods. By integrating minimally invasive delivery with established pacemaker technology, this strategy allows multi-site pacing, contrasting with conventional single-point methods. It could potentially facilitate planar wave propagation across ventricular tissue to normalize delayed myocardial conduction, offering a new way to mitigate re-entrant arrhythmias. Similarly, minimally invasive methods for the delivery of in vivo devices have been developed^9,145 (Fig. 6d). For example, catheter-like devices can be applied to deploy and attach silicon-based cardiac stimulators to pig hearts⁹. Once the delivery device is retracted, an optical fibre-coupled endoscope is inserted through the small incision for visually guided optical stimulation. This method was used to demonstrate the feasibility of non-genetic minimally invasive optical stimulation in a closed-thoracic setting.

Genetically targeted chemical assembly (GTCA) was developed to avoid the potential for tissue damage when interfacing with the intricate structures and neural networks of multicellular biological systems^146,147 (Fig. 6e). GTCA guides neurons to form in situ polymers with diverse electrical properties through cell-specific genetic triggers and a catalyst enzyme. For example, an ascorbate peroxidase, Apex2, can be genetically expressed on the surface of the neuron to trigger hydrogen peroxide-enabled oxidative polymerization of either conductive polymers or insulating polymers. Initial successes were tempered by non-specific membrane targeting of Apex2, which led to intracellular retention. Intracellular retention limits the permeation of large polymer precursors into living cells, leading to a low polymerization yield and increased cytotoxicity. Therefore, extracellular localization of Apex2 has been explored to increase yields, enhance biocompatibility, reduce potential toxicity and prevent interference with intracellular chemistry. For example, next-generation GTCA applies horseradish peroxidase optimized for cell surface reactions, resulting in densely clustered polymers around targeted neurons, with sustained cell viability and adaptability for anchoring various proteins, thus broadening the scope for future GTCA applications¹⁴⁸.

Soft conducting polymer gels use endogenous metabolites to trigger enzymatic polymerization within biological environments¹⁴⁹. This technique allows the in vivo formation of bioelectronics within the nervous system, offering a minimally invasive solution for nerve stimulation and direct interfacing with neural tissues. It represents a step towards integrating electronics with the dynamic nature of biological systems. These developments in bioelectronics aim to enhance therapeutic interventions and biological understanding by prioritizing minimal invasiveness and optimal integration with biological tissues¹⁵⁰.

Energy conversion beyond bioelectronic modulation

Energy conversion extends beyond the electrical and chemical modalities typically associated with bioelectronics. Methods harnessing alternative forms of energy, such as the conversion of light to heat in photodynamic therapy also have a role in medical treatments¹⁵¹. For example, photothermal materials¹⁵² and nanoparticles that absorb near-infrared light have been explored for non-invasive brain stimulation^36,153. Moreover, techniques including thermoelectric¹⁵⁴, magnetoelastic¹⁵⁵ and electrochemical¹¹⁷ energy harvesting are crucial to the operation of battery-free biological sensors, with applications such as sweat analysis¹⁵⁶.

Biomimetic concepts for energy conversion have driven innovations such as bioinspired synthetic nerves, which may enable light-driven energy conversion to facilitate wireless, tetherless signal transmission in soft bioelectronic devices^157–159. Such flexibility and adaptability would allow more effective interfacing with living tissues.

In addition, living systems are adept at energy conversion^160,161. Biological composites, that is, merging living cells with materials, are energy efficient and biocompatible^98,162,163. For example, leveraging living photosynthetic matter for the conversion of light into chemicals, such as oxygen, presents new avenues in tissue regeneration by aiding angiogenesis¹⁶⁴. The integration of living cells with optoelectronic devices may further allow direct stimulation of cardiomyocytes, among other applications¹⁶⁵. This symbiosis of living systems and energy conversion technologies holds potential for the development of new therapies and biomedical technologies⁹⁷ (Fig. 6f).

Citation diversity statement

We acknowledge that papers authored by scholars from historically excluded groups are systematically under-cited. Here, we have made every attempt to reference relevant papers in a manner that is equitable in terms of racial, ethnic, gender and geographical representation.

Key points.

• Implantable electroceutical devices often face challenges related to precision and biocompatibility.

• Indirect bioelectronic modulation approaches leverage energy conversion processes and show potential for precise, minimally invasive and effective therapeutic interventions.

• Photoelectrochemical and electrochemical principles can be used to convert electrical energy to or from alternative modalities for biological stimulation.

• There is a need to design advanced implantable electroceuticals with precise intracellular control, minimally invasive device deployment and living energy conversion systems.