Bioelectronic devices enable fundamental physiological and electrophysiological research, healthcare monitoring, and advanced therapeutics. To meet the demanding device requirements imposed by biomedical applications, graphene-based electronics offer a promising alternative to conventional bioelectronic device materials in an all-carbon platform. Continued advancements in graphene nanostructure synthesis and micro-fabrication techniques allow novel device architectures with vastly tunable physiochemical properties. Here, we highlight recent advances in graphene nanostructure-based bioelectronics. We distinguish between various material geometries and discuss their effect on device performance. Furthermore, we emphasize the continued development of fundamental relationships between 3D device geometries and material properties to allow next-generation bioelectronics for biosensing, electrophysiological recordings, and stimulation.
Bioelectronic devices enable real-time monitoring or control of physiological processes.1 Multidisciplinary progress in the fields of material science, microelectronics, and bioengineering has facilitated advancements in bioelectronics by pushing beyond existing limitations in material design and engineering. Fully realized bioelectronic technologies can facilitate medical diagnosis without clinical intervention or reverse detrimental effects of diseases.2–4 Devices that achieve stable, persistent functionality in soft tissue or under physiological conditions are of great interest in biological research as well as in healthcare.1,5–7 The functionality of these devices is governed by the properties of the interface between the bioelectronic materials and biological systems.8,9 For biosensors, such as glucose monitors, the sensitivity and limit of detection (LOD) of the device largely depend on the effective surface area of the electrode material at the interface.10–13 While, in the case of bioelectronic devices interfaced with electrogenic cells, the electrophysiology recording and stimulation capabilities of the devices are dependent on the impedance of the electrode,8,14 the exposed surface area,15 and the cell-electrode coupling.14 To enhance the overall performance of the devices, the physical and chemical topography of the interface must be precisely controlled. A key parameter governing the physical topography of the interface is the dimensionality of the sensing material.9
Conventional bioelectronics heavily rely on microfabricated two-dimensional (2D) thin films as sensing and actuating entities. Metals (e.g., gold) and semiconductors (e.g., silicon) have been instrumental in the development of bioelectronic devices.1,14,16,17 More recent biomimetic and bioinspired designs seek to match the biomechanics and architecture of the living tissue to achieve seamless integration of devices and tissues.4 Some examples include mesh electrode designs,18 flexible fibers,19 and thin polymeric membranes.20 While these efforts have contributed toward significant technological advances, the planar topographies of the fabricated interfaces exhibit inherent performance limitations that are a direct consequence of their limited exposed surface area and poor interactions with interfaced cells and tissues [Fig. 1(b)].8,9 Extending the dimensionality of the material design to a three-dimensional (3D) topography addresses this limitation by increasing the effective surface area exposed to the environment, thereby increasing the interaction of the analyte solution or cell(s) with the electrode surface. For biosensors, an increase in the effective surface area is directly correlated with an enhancement in the sensitivity and LOD.10–13 In the case of electrophysiology recorders and actuators, a non-planar topography increases cell adhesion, enhancing the electrical coupling between the cell membrane and the electrode.9,14 In addition, the expanded surface area of the electrode increases the overall electrode capacitance, reducing the effective impedance.8,21 Lowering the electrode impedance increases the signal-to-noise ratio (SNR) of the recorded electrical signals, enabling devices with a higher sensitivity and smaller geometric footprint.14 Nanomaterials with tailored physiochemical properties, shape, and dimensions provide a promising route to achieve these design criteria and allow seamless integration with biological systems.
Graphene, a honeycomb sp2 hybridized 2D carbon lattice, is one of the most recently discovered allotrope of carbon.22 It acts as the basic building block for other allotropes of carbon since it can be wrapped to form fullerenes, rolled-up to form nanotubes, and stacked to form graphite [Fig. 1(a-I)].23,24 Single crystal graphene exhibits advantageous properties such as chemical stability,25 electrical conductivity (charge carrier mobility up to 200 000 cm2 V−1 s−1),26 mechanical robustness (Young’s modulus of ∼1 TPa),27 a high surface-to-volume ratio (theoretical value of ∼2630 m2 g−1),28 and optical transparency (optical transmittance of ∼97.7%) [Fig. 1(a-II)].29
Functional graphene-based devices are enabled by graphene nanostructures composed of multiple graphene flakes following hierarchical assembly. Multiple schemes have been developed to obtain graphene nanostructures with hierarchical assembly resulting in application-specific properties. Bottom-up synthesis approaches, such as chemical vapor deposition (CVD)30,31 and pyrolysis,32 are based on the formation of graphene nanostructures through carbon-based precursor molecules or powders [Fig. 1(a-III)]. Such synthesis protocols take advantage of the self-assembling properties of aromatic structures at a molecular resolution, although these methods can be limited by regular or random structures and high defect densities.33 Alternative top-down approaches such as chemical exfoliation of graphite,34,35 and more recently, the laser-induced graphene (LIG) synthesis36,37 allow high yield rates; however, both present challenges for controlling resultant microstructures and defect populations.
For stable bioelectronic device performance, especially in the case of bioelectronics interfacing with cells and tissues, it is essential to evaluate the biocompatibility of the interfacing materials. The biocompatibility of graphene has been demonstrated;38 although careful considerations of the surface charge, chemical reactivity, size distribution, and shape for each intended biological application must be made.8,39 The multifaceted question of graphene biocompatibility has been reviewed extensively elsewhere.8,39,40 Another key aspect of device performance is the ability to chemically modify the material interface or add functional ligands [Fig. 1(a-IV)]. Graphene can undergo facile modification through non-covalent interactions between the graphene lattice and the aromatic ring(s) of the functionalizing molecule via π–π stacking.41,42 Graphene-based materials that express carboxylic groups at flake edges or hydroxyl groups at the basal plane can be functionalized via amidation or esterification, respectively.43 These chemical modifications can notably introduce biorecognition elements such as aptamers44,45 or antibodies46,47 for selective biochemical sensing.
With its exceptional physical and chemical properties, graphene has emerged as a promising material for bioelectronics. In this review, we rationalize the advantages of 2D and 3D topographies of graphene nanostructures by presenting current research trends in the domains of biochemical sensing, electrical recording, and stimulation. We highlight paradigms in the graphene nanostructure design that leverage on resultant structural, electrical, electrochemical, and optical properties for a multitude of application-tailored bioelectronics.
GRAPHENE-BASED BIOSENSING PLATFORMS
Current biochemical assays such as the enzyme-linked immunosorbent assay often require expensive analytical instruments and are unable to continuously monitor biomarker levels in tissues or patients. Electronics that incorporate biochemical sensors can provide sensitive, point-of-care detection tools to researchers and clinicians.48,49 To this end, graphene-based biochemical sensors have made significant strides recently by demonstrating selective detection in complex biological fluids,50–53 high spatiotemporal resolution,54–56 and translation to in vivo platforms.55–59 This section tracks recent works that incorporate graphene nanostructures into biochemical sensing platforms to improve diagnostic tools and further clinical applications in the field.
Researchers have demonstrated the incorporation of graphene nanostructures into existing electrochemical sensing platforms to improve device sensitivity and decrease the LOD.10,13,53,60 Cardoso et al. demonstrated laser-induced graphene (LIG)-based sensors for the detection of chloramphenicol (CAP) in a standardized, three-electrode design [Fig. 2(a-I)].10 The authors patterned their devices on a flexible platform with a molecularly imprinted polymer (MIP) layer specific to CAP [Fig. 2(a-II)].10 SEM micrographs reveal 3D porous graphene structures [Fig. 2(a-III)] at the electrode surface. The LIG biosensors showed improved performance over commercial carbon screen-printed electrodes and demonstrated a LOD of 0.62 nM with a wide linear response range from 1 nM to 10 mM [Fig. 2(a-IV)]. Graphene structures have also been readily integrated into standard 3D printing methods for the rapid development of biosensors.53,61 Marzo et al. fabricated 3D graphene/PLA electrodes for H2O2 detection by fused deposition modeling. The fabricated sensors retained 98% and 84% of their measured sensitivity to H2O2 in buffer solution when introduced to 1:4 diluted human serum and five days after the introduction of the serum, respectively.53
Notably, researchers are also incorporating graphene-based biochemical sensors into in vivo platforms.55–59 Liu et al. developed an implantable microelectrode array (MEA) for in vivo monitoring of neuro-chemical and electrical signals using a reduced graphene oxide–anodized gold (rGO/Au2O3) nanocomposite electrode [Figs. 2(b-I) and 2(b-II)].55 Functionalizing the gold electrodes with rGO resulted in a 108% increase in the probe sensitivity toward hydrogen peroxide (H2O2) [Fig. 2(b-III)]. The coated electrodes showed a rapid response to the increasing H2O2 concentration in the hypoxic rat brain cortex [Fig. 2(b-IV)]. The authors couple their electrochemical measurements with recordings of somatosensory-evoked potentials to further assess functional changes of neuronal activities during induced stroke. Kim et al. incorporated transparent graphene/Ag nanowire (graphene/AgNW) nanostructures into wearable contact lenses for glucose detection and intraocular pressure measurements.57 The conductive graphene/AgNW composites form source and drain current contacts while graphene channels modified with glucose oxidase allow for selective field effect-based sensing of glucose. The contact lens devices achieved a LOD of 1 μM, a ten times improvement in LOD over previously reported contact lens sensors formed with evaporated metal electrodes.
Recently, functionalized graphene field effect transistors (gFETs) have shown remarkably low detection limits for trace biomarkers44,50,54,62,63 due to FET’s signal transduction mechanism and unique operating principles.64,65 Hwang et al. significantly improve gFET device performance for sensitive nucleic acid detection via crumpled graphene channels [Fig. 2(c)].50 The authors form gFET channels with and without substrate deformation that results in a wrinkled channel. DNA is immobilized on the graphene channel through a pyrenebutanoic acid succinimidyl ester (PASE) linker on the channel, and the selective detection is enabled by the hybridization of the target DNA with a probe strand of DNA. The authors compare the sensitivity of gFET devices for the detection of miRNA (let-7b) with and without deformed graphene channels, where crumpled gFETs show LODs of ∼20 aM in human serum. Authors posited the formation of electrically charged regions in the nanoscale crumpled graphene channel with an increased Debye length, resulting in significant Dirac potential shifts due to the charge of DNA/RNA. Notably, Wu et al. fabricated dual-aptamer modified gFET biosensors for label-free detection of hepatocellular carcinoma (HCC)-derived micro-vesicles (HepG2-MVs) in clinical blood samples.62 Gold nanoparticles with both HepG2 cell specific TLS11a aptamer (AptTLS11a) and epithelial cell adhesion molecule aptamer (AptEpCAM) were bound to the gFET channel for the capture and quantification of HepG2-MVs. The authors find significant differences in HepG2-MVs quantities between healthy control groups and HCC patients using their gFET biosensing platforms, indicating a promising potential for early HCC diagnosis.
Graphene nanostructures clearly provide conductive, high surface area electrode interfaces that can be readily functionalized or composited with biorecognition elements to provide high selectivity. Graphene-based sensors that demonstrate 3D topologies are expected to achieve faster response times, lower LODs, wider linear response ranges, and greater sensitivity values. Additional studies are needed to further clarify relationships between specific 3D geometries and biosensor performance.
GRAPHENE-BASED ELECTROPHYSIOLOGY RECORDING PLATFORMS
Various graphene-based platforms have been developed for recording electrical activity from individual cells as well as tissues in vitro and in vivo.66–68 The most common 2D graphene electrode assemblies utilize the graphene either as a passive electrode, i.e., an electrode material in multi-electrode arrays (MEAs),68–70 or as an active electrode, i.e., a semiconducting channel in FETs.66,67 In addition, the mechanical properties of graphene allow the fabrication of bioelectronics on both rigid as well as flexible substrates.69,71
The high transparency of 2D graphene allows simultaneous optical, e.g., monitoring Ca2+ signaling using fluorescent dyes, and electrophysiological recordings from 2D graphene-based electrodes.21,68,69 Rastogi et al. recently demonstrated in vitro electrical recording of extracellular field potentials and corresponding Ca2+ activity of cardiomyocytes using 2D graphene based MEAs [Fig. 3(a)].70 Differential interference contrast (DIC) images of cardiomyocytes cultured directly on the MEA platform demonstrate the optical transparency of the MEAs that enable Ca2+ imaging [Fig. 3(a-I)].70 The MEA platform demonstrated stable electrical recordings of extracellular field potentials with high spatial and temporal resolution with a signal-to-noise-ratio (SNR) greater than 14.70 The optical transparency of 2D graphene also allows graphene-based transistor arrays to be supplemented with scanning photocurrent microscopy to measure the electrical activities of individual synapses of primary hippocampal neurons [Fig. 3(b)].72 The high conductivity of graphene generates an ultrafast photocurrent response in the transistors leading to sub-millisecond temporal resolution.72 This was demonstrated by chemically stimulating neuronal bursts by increasing the extracellular K+ concentration from 4 mM to 60 mM.72
A bottleneck for integrating high density transparent 2D graphene sensors on to a single platform is that conventional graphene sensors require each sensing element to be individually connected to a signal amplifier.73 Garcia-Cortadella et al. have integrated graphene transistors with custom-built front-end amplifiers to demonstrate the concept of frequency-division multiplexing of neural signals.73 This was achieved by performing on-site amplitude modulation of the recorded neural signals with graphene transistors.73 The 4 × 8 array of graphene solution-gated FETs (g-SGFETs) was interfaced with the cortex of the Long Evans rat to record the spatial distribution map of cortical spreading depression (CSD) events [Fig. 3(c)].73 The presented approach eliminates the need of switches avoiding the limitations due to limited switching speeds simplifying the overall complexity of the platform.
Recording electrical activity with sub-cellular precision using graphene-based electrodes requires the fabrication of ultra-microelectrodes.19 To achieve this, the geometric footprint of the individual electrodes needs to be reduced. However, this leads to an increase in the overall impedance of the electrode, degrading the recording capabilities of the platform.14,19,74 Various approaches have been developed to decrease the impedance of UMEs including deposition of surface coatings, such as Pt black and conductive polymers.75,76 Delamination of surface coatings reduces the long-term stability of graphene bioelectronic interfaces.77 Extending the topography of 2D graphene to 3D allows the effective surface area of the material to be drastically increased.31,78 This was demonstrated by the highly controlled out-of-plane synthesis of single-to-few-layer graphene flakes on a Si nanowire (NW) template: nanowire templated 3D fuzzy graphene (NT-3DFG).31 The 3D arrangement of out-of-plane graphene flakes was shown to increases the electrical conductivity79 as well as the effective electrochemically active surface area of the electrodes.31,80
Rastogi et al. fabricated NT-3DFG-based UMEs as a novel platform for recording extracellular field potentials from human embryonic stem cells-derived cardiomyocytes (hESC-CMs) without the need for any additional surface coatings [Figs. 4(a) and 4(b)].81 The increase in the effective surface area of the electrodes due to the 3D out-of-plane arrangement of graphene flakes drastically reduces the impedance of NT-3DFG electrodes with a geometric footprint of 50 × 50 μm2 to 9.4 ± 2.7 kΩ as compared to 1.20 ± 0.16 MΩ for Au-based electrodes with the same geometric footprint [Fig. 4(a)].81 This allows the fabrication of NT-3DFG UMEs with the geometric footprint as small as 10 × 10 μm2 and 5 × 5 μm2 while still exhibiting impedance significantly lower than 50 × 50 μm2 Au (236 ± 22 kΩ and 447 ± 72 kΩ, respectively).81 Downsizing NT-3DFG MEAs to 2 × 2 μm2 led to electrodes with impedance of 1.23 ± 0.18 MΩ.81 In addition, the novel topography of graphene enabled the creation of stable biointerfaces with enhanced coupling with cell membranes without inducing cell stress.81 NT-3DFG UMEs were employed to record extracellular field potentials of hESC-CMs with an amplitude of 400 µV–800 µV and a SNR greater than 6 [Fig. 4(b-IV)].81 This platform exhibits high temporal resolution for electrical recordings and allows deconvolution of signals pertaining to the Na+ (upstroke), Ca2+ (plateau phase), and K+ (repolarization) currents across the cell membrane.81 Other 3D graphene based passive electrodes including recording platforms are fabricated using graphene fiber microelectrodes82 and porous graphene electrodes.21
The minimum detectable signal using graphene-based transistors is inversely proportional to the square-root of the electrochemically active surface area of the transistor channel.83 An attractive approach to addressing this limitation is to design highly crumpled graphene-based transistors [Fig. 4(c-I)].84 The 3D topography of the graphene channel can be obtained by subjecting large biaxial compressive forces to form a network of high aspect-ratio crumples.84 The high fracture strain of graphene prevents structural failure under the applied compressive forces and material delamination.78,84 Highly crumpled graphene based transistors have demonstrated stable real-time recording of the electrocorticography (ECoG) from the left cortex of an anesthetized rat [Fig. 4(c-II)]. The temporal resolution of the ECoG activity recordings allows the identification of the basal activity, latent period, and induced epileptiform activity periods with high SNR.84
The dimensionality of graphene nanostructures greatly influences the optical and electrochemical properties of the material. For concurrent electrical and optical electrophysiology recordings, 2D graphene nanostructures are preferred. However, for a reduced sensor footprint and enhanced detection limits, 3D graphene nanostructures are preferred as they provide efficient electrical coupling and lower electrode impedance.
GRAPHENE-BASED STIMULATION PLATFORMS
The ability to modulate the electrophysiology of cells and tissues helps develop a deeper understanding of inherent developmental and functional processes of healthy and diseased entities. Electrical stimulation of central and peripheral nervous systems has been clinically demonstrated to reduce tremors and motor rigidity in patients with neurological disorders such as Parkinson’s disease.85 This is commonly achieved through capacitive charge injection from the electrode surface to the target entity through charging and discharging of the electrode–electrolyte double layer.15 Various graphene-based platforms have been developed for electrical stimulation of in vitro and in vivo systems to understand the fundamental charge transfer processes from graphene electrodes to target tissues as well as develop therapeutic tools.21,82,86–88
2D graphene-based electrodes have been demonstrated to be effective electrical stimulators.86,87 The charge injection capacity (CIC) of a stimulating electrode is a measure of the total charge per unit area delivered to the target entity in the leading phase of the stimulation pulse.15 Park et al. estimated the CIC for fabricated graphene electrodes [Fig. 5(a-I)] to be 57.13 μC/cm2 at a negative polarization potential of −0.6 V [Fig. 5(a-II)].86 The high optical transparency of 2D graphene enables concurrent spatiotemporal imaging of neural response for understanding the mechanisms of electrical stimulation.86 Temporal mapping of the cellular activity with fluorescent indicators reveals that the change in fluorescence intensity increased as the amplitude of the stimulation pulse was increased [Fig. 5(a-III)]. This suggests a direct correlation between neural activation and the injected charge.86
As discussed earlier, the 3D topography of the electrode increases the effective surface area, resulting in an increased charge storage capacity.15 To this end, Wang et al. reported the fabrication of free standing high-performance Pt-coated graphene-fiber (GF) based MEAs.82 The GF electrodes are stiff enough to directly penetrate soft nerve tissues and flexible enough to avoid mechanical mismatch with the interfaced tissues [Fig. 5(b-I)]. The porous structure of GF drastically enhances the CIC of the electrodes to 10.34 mC/cm2. These electrodes demonstrate stable electrochemical performance over multiple cyclic voltammetry cycles and charge injection pulses [Figs. 5(b-II) and 5(b-III)].82
An alternate approach to modulating the electrical activity of neurons is through remote nongenetic photothermal stimulation using the graphene nanostructures of NT-3DFG [Fig. 5(c)].89 The unique assembly of out-of-plane graphene flakes causes the material to exhibit high broadband optical absorbance (∼95%) and efficient photothermal energy conversion with NT-3DFG exhibiting ∼140- to ∼260-fold greater increase in local temperature change compared to Si nanowires [Fig. 5(c-I)].89 NT-3DFG adheres to the plasma membrane of cells rather than being internalized [Fig. 5(c-II)]. This interfacial coupling enables photothermal stimulation of neurons via the optocapacitive mechanism, where upon illumination, the localized heating of the NT-3DFG and cell membrane generates a depolarizing capacitive current.89 By increasing the power of the illuminating pulse, subthreshold excitations are converted to action potentials [Fig. 5(c-III)]. Photothermal stimulation with NT-3DFG is highly reproducible, can be used to apply pulsed stimulation of the target neurons at varying frequencies, and does not induce damage to the cell membrane under routine stimulation conditions.89 Compared to the size of a neuronal body, NT-3DFG has a significantly smaller diameter that enables sub-cellular spatial resolution and real-time Ca2+ imaging to track signal transduction across cellular networks in 2D cell cultures and 3D spheroids.89
Compared to 2D graphene nanostructures, 3D topography results in efficient electrical stimulation platforms due to the increased effective surface area and higher density of exposed edge planes. Moreover, the increased optical absorption by 3D nano-topographies opens opportunities for remote nongenetic optical stimulation paradigms.
CHALLENGES AND PROSPECTS
While large developmental strides have been made for graphene-based bioelectronics, many multidisciplinary challenges remain for clinical applications or widespread adoption. Often, the device lifetime and operational stability are key concerns for research applications or clinical translation.3,4 To address this challenge, steps must be taken to ensure both the chemical and structural integrity of the electrode material and prevent degradation of the device interface. Recent studies have shown the reliability of graphene microelectrodes on flexible PET substrates for neural stimulation under accelerated aging conditions for over one month.90 For 3D nanostructures, porosity must be balanced against desired mechanical stiffness, and structures should be bound to the device substrate to prevent delamination or loss of material. Compositing graphene nanomaterials with porous polymers can provide additional structural stability or bind nanomaterials to device substrates.56,91 For in vitro and in vivo applications, the effects of biofouling through non-specific protein adsorption at electrode surfaces should be considered.92,93 Notably, carbon-based materials can be readily functionalized with antifouling coatings,93,94 though the resultant effects on the device response time or sensitivity should be taken into consideration.95
In terms of device fabrication, the varied synthesis strategies for graphene-based nanomaterials allow for many possible processing pathways involving suspensions,96 composites,34,91 or thin films.97 Nevertheless, practical approaches must involve the integration of graphene nanostructures with the passive components of device platforms. Robust methods that allow the transfer of graphene or direct synthesis of graphene nanostructures onto source fabrication wafers, for example, are preferable for integration with established microfabrication techniques.81,98,99 3D topographies can add complexity to device fabrication; however, sacrificial polymer layers or hard masks can be implemented to protect 3D topographies during further processing steps.81,100
The topography of the bioelectronic interface and the dimensionality of the constituent materials play a dominant role in the overall device performance. Here, we have highlighted recent advancements in graphene-based nanostructures toward bioelectronics for biochemical sensing, electrophysiology recording, and stimulation. 2D graphene enables simultaneous optical and electrical recordings of electrophysiology. Furthermore, 3D topography drastically increases the exposed surface area resulting in increased interactions with cells and enhanced electrochemical properties. Truly 3D graphene topographies based on free-standing graphene flakes exhibit an exceptionally high surface area and superior electrochemical properties by exposing graphene flakes’ electrochemically active edges and both sides of the basal plane. This has further pushed the physical dimensions of bioelectronics to sub-cellular regimes. Continued advancement in 3D graphene bioelectronics will require the investigation of the complex relationships between 3D arrangements of graphene flakes and the resultant device characteristics. These efforts will enable high-performance, application-tailored graphene nanostructure-based devices for the next generation of bioelectronics.
D.S.R. and R.G. contributed equally to this work.
T.C.K. acknowledges funding support from the NSF under Award No. CBET1552833, the Office of Naval Research under Award No. N000141712368, and the Defense Advanced Research Projects Agency under Award No. AWD00001593 (416052-5).
Data sharing is not applicable to this article as no new data were created or analyzed in this study.