Application of Graphene in Neural Activity Recording

Suye Lv; Liang Zou; Shouliang Guan; Hongbian Li

doi:10.7536/PC200811

Progress in Chemistry ›› 2021, Vol. 33 ›› Issue (4) : 568-580 DOI: 10.7536/PC200811

review-article

Scan to read & Share

Application of Graphene in Neural Activity Recording

Suye Lv ¹^,²^,³
, Liang Zou ¹^,²^,³
, Shouliang Guan ¹^,²^,³
, Hongbian Li ¹^,²^,^*

Author information +

History +

PDF

Abstract

As a powerful tool for monitoring brain activities, neural electrodes have been playing a crucial role in the understanding of brain functions and the treatment of neurological disorders. In particular, the construction of a stable electrode-neural interface is critical for the stable chronic neural recording. However, conventional neural electrodes are mainly constructed with rigid materials, whose Young’s moduli are several orders of magnitude higher than that of the brain tissue. This large mechanical mismatch causes the micromotion of the neural electrodes, which elicits inflammatory response of the brain tissue and thus limits the stable chronic neural recording. With the one-atom thickness, graphene has been considered as a promising active material for neural electrodes with stable electrode-neural interfaces for its high conductivity, excellent flexibility and good biocompatibility. In this review, we provide an overview of graphene for neural activity recording, from their modulation on the growth of neurons to applications in bothin vitro and in vivo neural activity recording. At last, challenges and prospects of graphene for neural activity recording are proposed.

Contents

1 Introduction

2 Graphene microelectrodes for cell culture and in vitro neural activity recording

2.1 Graphene for cell growth and modulation

2.2 Graphene microelectrodes for in vitro neural activity recording

3 Graphene microelectrodes for in vivo neural activity recording

3.1 Graphene-based electrocorticography(ECoG) electrodes for neural activity recording

3.2 Graphene-based intracortical electrodes for neural activity recording

4 Conclusion and prospects

Graphical abstract

Keywords

graphene / neural electrodes / electrophysiology / interface / stable chronic recording

Cite this article

Download citation ▾

Suye Lv, Liang Zou, Shouliang Guan, Hongbian Li. Application of Graphene in Neural Activity Recording. Progress in Chemistry, 2021, 33(4): 568-580 DOI:10.7536/PC200811

登录浏览全文

4963

注册一个新账户忘记密码

1 Introduction

The functional realization of human brain depends on the coordinated electrophysiological activity of a large number of neurons^[1]. It is of great significance to improve the understanding of the mechanism of neural circuits and the treatment of neurological diseases (such as Parkinson's syndrome, epilepsy, etc.) By using neural electrodes to record the electrophysiological activities of neurons and to stimulate neurons to regulate neural electrical activity^{[2⇓⇓⇓~6]}. In addition, the establishment of brain-computer interface between neurons and external devices through neural electrodes, the development of a new generation of neuroprosthetic devices for the control of prostheses and motor rehabilitation, and the realization of the control of externally driven prostheses provide a new opportunity for the repair or replacement of damaged neural functions^[7,8]. To achieve highly sensitive and stable measurement of neural electrical signals and efficient and stable stimulation of neurons, the construction of a reliable electrode-nerve interface is the key^[9]. After nearly 70 years of development, neural electrodes, such as Utah electrode, Michigan electrode and metal microwire electrode, have been able to record and stimulate the electrical activity of neurons with high spatial and temporal resolution^[10⇓~12]^[13,14]^[15]. At the same time, based on the rapid development of neural electrodes, the establishment of brain-computer interface has enabled patients with motor and sensory loss to break through their own mobility barriers and communicate with the outside world^[3,16]. However, at present, neural electrodes are mainly composed of rigid materials, such as silicon and noble metals (gold, platinum, iridium). The Young's modulus of these materials is about 100 ~ 500 GPa, while that of brain tissue is only 100 Pa ~ 10 kPa^{[10,13,17⇓⇓⇓~21]}^[22]. The huge difference in mechanical properties makes the traditional nerve electrode implanted into the brain move slightly with the head movement, heartbeat, breathing and other physiological activities, which further induces the accumulation of glial cells and the formation of glial scar^[23⇓~25]. As time goes on, glial cells act as an insulating layer on the surface of the nerve electrode, thus hindering the long-term stable recording of the electrophysiological signals of neurons by the electrode^[26,27]. In addition, the corrosion of metal electrodes greatly limits the application and development of traditional nerve electrodes^[28⇓~30].

Graphene is a two-dimensional nanomaterial composed of sp² hybridized carbon atoms with a single atomic layer thickness^[31]. As an active material for neural electrodes, graphene has many advantages that other materials do not have: (1) Graphene has good mechanical flexibility, which can form close contact with soft brain tissue to build a stable electrode-nerve interface^[32]. (2) Graphene has both good electrical conductivity and excellent carrier mobility, and its carrier mobility is as high as 100 000 cm²·V^-1·s^-1[33], which provides conditions for highly sensitive detection of neural electrical signals by neural electrodes based on graphene field effect transistors. (3) Due to the thickness of only one atomic layer, graphene has a very high specific surface area, combined with its unique electrical properties, which makes the graphene neural electrode have low electrochemical impedance and high charge injection ability, laying the foundation for its effective electrical stimulation of neural tissue^[34]. (4) Graphene has high transparency, and the transmittance of single-layer graphene is as high as 97.3%^[35]. Therefore, graphene electrodes can be used to record electrophysiological signals while observing the changes of cells and blood vessels under the electrodes with electrophysiological activities, which provides a technical means for a clearer understanding of brain structure and function. In addition, based on the high transparency, graphene neural electrodes can also be combined with photogenetics and other technologies to achieve optical regulation of neural electrical activity. Based on the above advantages, graphene has great potential for development in basic neuroscience and brain-related diseases.

In this paper, the research progress of graphene in neural electrophysiological detection is reviewed, including the regulation of graphene on nerve cells and the recording of neural electrical signals in vitro and in vivo using graphene electrodes. Finally, the future development of graphene neural electrodes is prospected.

2 Interaction of Graphene with Cells and Detection of Neural Electrical Signals in Vitro

Nerve cell is an important unit of brain tissue, and the generation of neural electrical signals is mainly based on the electrophysiological activities of nerve cells. During the electrophysiological activity of cells, a large number of K⁺ and Na⁺ pass through the ionic transmembrane and cause the change of extracellular potential^[36]. Nerve electrode can be used to understand the activity of nerve cells and the mechanism of pathological changes by recording the changes of extracellular potential of nerve cells. Using neural electrodes to monitor the electrical signals of cells or tissues in vitro can avoid the invasive damage caused by neural electrodes to the living body and provide guidance for in vivo experiments. In the following, we will introduce the modulation of graphene on cultured nerve cells and its application in the detection of neural electrical signals from two aspects: the interaction between graphene and nerve cells and the detection of neural electrical signals in vitro using graphene electrodes.

2.1 Interaction between Graphene and Neural Cell

graphene has good chemical stability and biocompatibility,so it can provide an ideal platform for the culture of nerve cells.In addition,graphene can also regulate the growth,differentiation and proliferation of cells cultured on its surface.on this basis,we can further regulate the electrical properties of the neural network,thus helping to rebuild the damaged central nervous circuit and achieve nerve injury repair.the graphene has a hydrophobic structure,and hydrophilic molecules such as polyethylene glycol,polylysine,polyacrylic acid and the like are modified on the surface of the graphene,so that the attachment of cells on the surface of the graphene can be enhanced,and the survival and growth of nerve cells on a graphene substrate can be promoted^[37⇓~39]^[40]^[41]。 For example,Li et al used a graphene film prepared by chemical vapor deposition as a culture substrate,and further coated it with poly-lysine(PLL)to enhance cell attachment^[42]。 Compared with the traditional polystyrene substrate,the Cultured mouse hippocampal neurons had faster synaptic sprouting and growth rate at the early stage of culture.In addition,Agarwal et al.cultured rat adrenal pheochromocytoma(PC12)cells,oligodendrocytes and osteoblasts on single-walled carbon nanotube networks and graphene films,respectively^[43]。 the results showed that PC12 cells,oligodendrocytes and osteoblasts could proliferate well on the graphene film,but the proliferation and activity of cells were inhibited on the carbon nanotube network。

in recent years,neural stem cells(NSCs),as a pluripotent cell population with self-renewal ability in the central nervous system,have attracted wide attention.neural stem cells have two main characteristics:(1)They have unlimited self-renewal capacity.(2)pluripotency,the ability to differentiate into all types of cells in the neuroectodermal lineage of the central nervous system,such as glial cells and various nerve cells.Using neural stem cells for brain repair and nerve regeneration is the key to the treatment of a variety of neurological diseases.the differentiation of NSCs is affected By both extracellular and intracellular factors,especially the specific microenvironment and the metabolic state of the cells.by modifying collagen,laminin and fibronectin on the surface of graphene and graphene oxide,a biomimetic microenvironment can be constructed to promote the attachment,growth,proliferation and differentiation of human neural stem cells(hNSCs)^[44]^[45]^[46]^[47]。 for example,Park et al.Used laminin-modified graphene film as a substrate to culture human neural stem cells,and found that the graphene substrate could provide a more favorable microenvironment For human neural stem cells,thus promoting the differentiation of neural stem cells into neurons(as shown in Figure 1A)^[45]。 in addition,graphene can also promote the differentiation of human mesenchymal stem cells(hMSCs)into non-mesenchymal tissues,such as neurons,with the assistance of chemical,physical or mechanical stimuli.Wang et al.Used the strong polarity of carbon-fluorine bonds In fluorinated graphene to induce bone marrow mesenchymal stem cells to differentiate into neuronal lineage^[47]。 It is Found that graphene with strong polarity can induce the orientation of cell morphology,from irregular polygon to spindle,and the growth of nucleus,which is more conducive to the differentiation of neurons.in addition,the differentiation of mesenchymal stem cells into neuronal lineages can also be induced by patterning fluorinated graphene into aligned elongated microchannels without chemical inducers(Fig.1b,1C).Later,Kim et al.found that In the early stage of culturing hMSCs on highly uniform graphene substrates,hMSCs would temporarily form three-dimensional spherical shapes^[48]。 This three-dimensional sphere upregulated the expression of genes related to neurogenesis in hMSCs,thereby promoting the differentiation of hMSCs into neuronal lineages。

although amino acids,proteins and other modifications can effectively increase the attachment of cells to graphene,the introduction of these coatings also increases the impedance of neurons/interface,affects the charge transfer at the interface,and reduces the sensitivity and reliability of electrophysiological signal detection.Therefore,the direct culture of nerve cells on the graphene substrate can enhance the contact and electrical coupling between graphene and nerve cells,thus realizing the detection of weak electrical signals.Bendail et al.First demonstrated that primary retinal ganglion cells could survive on graphene substrates without any glial support layer or protein coating,Although the synapse length was reduced compared with that on graphene substrates with PLL coating^[49]。 Fabbro et al.Directly used graphene microsheets prepared by liquid phase exfoliation and ball milling as substrates to culture hippocampal neurons,and found that graphene substrates without protein coatings were inert neural interface materials,which could retain the electrophysiological characteristics of nerve cells and did not affect the charge transfer between nerve cells^[50]。 in addition,Veliev et al also found that the crystallinity of graphene plays an important role In the attachment,growth and axon formation of nerve cells^[51]。 High quality graphene is a necessary condition for the effective attachment of nerve cells to its surface.With the decrease of graphene crystal quality,nerve cells change from highly attached to completely repelled on its surface。

Nerve cells in organisms exist in a three-dimensional growth environment^[52]。 Compared with graphene film,three-dimensional graphene scaffolds can better simulate the real environment of organisms,such as cell-cell and cell-matrix interactions,so three-dimensional graphene scaffolds as cell culture carriers can better reflect the electrophysiological behavior of nerve cells in real organisms^[52]。 In addition,three-dimensional scaffolds can promote the proliferation and differentiation of PC-12 cells and neural stem cells^{[53⇓⇓~56]}。 Three-dimensional graphene foam and rGO fiber cell scaffolds can ensure the large-scale transport of nutrients needed for the metabolism of neural stem cells due to their nano-porous structure,so that neural stem cells are in an active state of proliferation and differentiation^[56]^[57]。 For example,Li et al.used three-dimensional graphene foam synthesized by chemical vapor deposition on nickel foam as a scaffold to culture hNSCs(as shown in Fig.1D,1e)^[54]。 It was found that three-dimensional graphene foam could significantly enhance the differentiation of hNSCs into nerve cells and glial cells.Due to its good electrical conductivity,graphene can also be used as an electrode to stimulate differentiated hNSCs.In addition,three-dimensional tubular scaffolds of graphene can also be used to create cell structures with highly precise geometric shapes,and brain-like functional tissues can be constructed by higher-order assembly of spatially arranged cell structures.Sakai et al.Proposed a graphene/polymer bilayer film that can self-fold to form microcoils for neural cell culture^[58]。 At the same time,holes were created in the microcoils,allowing the nerve cell axons to extend through the holes to form functional connections with the surrounding environment(fig.1f,1G).Qian et al.Further applied a three-dimensional graphene substrate to an artificial nerve conduit(as shown in Fig.1H)^[59]。 The catheter is formed by printing polydopamine/polypeptide and graphene/polycaprolactone layer by layer,wherein the polydopamine and the polypeptide can increase the affinity of graphene and cells,and the graphene and the polycaprolactone can reinforce a tubular structure to provide certain rigidity for long-term in-vivo research.the nerve conduit effectively promoted axonal and myelin regeneration in an animal model of sciatic nerve injury,resulting in peripheral nerve repair(as shown in Figure 1 I)。

neural network electrical signals refer to the electrophysiological cooperative activity of a population of neurons that constitute functional connections.Studying the electrophysiological activities of neuronal networks is the key to understand the formation of neural circuits,and it is of great significance to regulate the electrical signals of neural networks reasonably for the treatment of nervous system diseases.Tang et al.differentiated hNSCs on the graphene substrate and found that the Differentiated hNSCs could grow on the graphene surface and form a functionally connected network^[60]。 in addition,the excitability and activity of the neural network on the graphene substrate were significantly enhanced,which proved the feasibility of graphene to regulate the behavior of neural network In vitro.Later,Pampaloni et al.Found that single-layer graphene can restrict the migration of potassium ions near its surface,thereby regulating neuronal communication and enhancing membrane ion currents^[61]。 It was found that the frequency of neuronal synaptic currents recorded on graphene substrates was significantly higher than that of the control group(as shown in Figure 1 J)。

the above results show that graphene has good compatibility with nerve cells.In addition,graphene can also effectively regulate the growth behavior of nerve cells and the electrical signals of neural networks.Based on these properties,graphene is expected to be used to develop neural interfaces that are friendly to the nervous system,creating opportunities for the preparation of new neural electrodes。

2.2 Detection of neural electrical signals in vitro with graphene neural electrode

graphene has excellent biocompatibility and electrical properties,which can achieve effective functional interaction with neural tissue,making it possible to regulate and record neural electrophysiological activities In a wide range,with high spatial and temporal resolution and long-term stability.in addition,compared with traditional metal electrodes,graphene has extremely high transparency,which can be transferred to transparent substrates to prepare transparent nerve electrode arrays,which can record the electrophysiological signals of nerve cells without affecting the imaging below the site.At present,graphene-based liquid gate transistors and electrode arrays have been widely used for extracellular action potential measurement of cardiomyocytes and cardiomyocyte-like cells^{[62⇓⇓⇓~66]}^[67,68]。 Compared with cardiomyocytes and cardiomyocyte-like cells,the electrophysiological activities of cortical neural networks are more complex and random.Therefore,the electrical properties of the electrode active material are required to be higher.Chen et al.Realized the recording of crayfish axonal action potential by reducing the impedance of graphene by vapor plasma treatment on the surface of graphene^[69,70]。 This electrode has only one very large recording site,which limits the resolution of the electrode recording.Du et al.Prepared a graphene microelectrode array on a quartz substrate to record the electrophysiological activity of cortical neurons isolated from rat embryos(as shown in Fig.2a,2b)^[71]。 The signal-to-noise ratio of the device is 10.31±1.2(see Figure 2 C).Later,Kireev et al.Made similar graphene microelectrode arrays on SiO₂/Si and boride glass substrates^[72]。 the signal to noise ratio(SNR)of the device was(48±26)when recording the electrophysiological signals of the embryonic cortical neurons of rats.Compared with planar electrodes,three-dimensional electrodes can be better coupled with cells and have more advantages in extracellular potential measurement.Jeong et al.developed a neural electrode based on a hybrid material of Graphene and carbon nanotube arrays(Graphene-VACNT)(as shown in Fig.2D)^[73]。 Vertical carbon nanotubes can form a closer contact interface with cultured nerve cells(as shown in fig.2e).Therefore,this electrode shows a higher signal to noise ratio than a planar TiN electrode(as shown in fig.2F)。

In order to realize the detection and recording of the electrophysiological signal of a single neuron by the graphene microelectrode,the size of the graphene microelectrode should be equivalent to that of a single neuron.However,as the size of the microelectrode decreases,its impedance increases,the thermal noise of the device increases,and the signal-to-noise ratio decreases.Compared with microelectrode arrays,transistor devices have the advantages of signal self-amplification,low noise and high sensitivity.In addition,its transconductance depends on the geometric shape factor(aspect ratio)of the channel,which breaks through the limitation of the size effect of the microelectrode array and makes it easier to construct neural electrodes with high spatial resolution.Cheng et al.First recorded the spontaneous discharge signals of cerebral cortical neurons cultured on the surface of graphene liquid-gate transistor(SGFET)electrodes^[74]。 SGFETs exhibit higher transconductance and lower noise levels than comparable transistor devices.In addition,the device has good mechanical flexibility,and its performance remains unchanged after repeated bending,which lays a foundation for the preparation of flexible nerve electrodes for electrophysiological signal detection in vivo.Thereafter,Veliev et al.fabricated graphene liquid gate transistor devices on flexible polyimide substrates and transparent substrates(sapphire,glass,SiO₂/Si)respectively(as shown in Fig.2g),and used them for neural signal recording of primary hippocampal neurons in mice(as shown in Fig.2h~2j )^[75]。 Experiments show that the sensitivity of graphene liquid gate transistor with sapphire substrate is the highest,which may be related to the lower charge density bound at the interface of sapphire substrate.in addition,In order to build a more stable graphene-neuron interface,Kireev et al.Developed a new packaging method for graphene liquid gate transistors^[76]。 By encapsulating only the metal wires on the chip,the cultured neurons are prevented from generating bending stress when approaching the graphene,thereby reducing the gap between the neurons and the graphene.Because of its single atomic layer thickness,graphene often needs to be transferred from the growth substrate to the device substrate with the help of a polymer support layer.This process will introduce polymer residues on the surface of graphene,thus increasing the contact resistance between graphene and metal electrodes,resulting in high contact noise.Schaefer et al.Used ultraviolet ozone to treat graphene to remove the residual glue on the surface of graphene and reduce the contact resistance between graphene and metal^[77]。 the results show that the graphene devices treated with ozone show lower contact noise and higher signal-to-noise ratio in the periodically generated hippocampal slice signal recordings than the untreated devices。

the above results show that graphene shows excellent electrical properties and biocompatibility in neural signal recording in vitro.Through surface modification or hybridization,the impedance of the graphene and the biological interface can be further reduced,thereby improving the signal-to-noise ratio of the recorded electrophysiological signal and laying a foundation for recording the electrophysiological signal in vivo。

3 Recording of nerve electrophysiological signals in vivo with graphene nerve electrode

Compared with the neural electrophysiological signals in vitro, the detection of neural electrophysiological signals in vivo can be less disturbed by the external culture environment, and can better reflect the real activity of neurons in brain tissue. Graphene has good mechanical flexibility and biocompatibility, as well as excellent electrical properties, which can meet the long-term stable recording and electrical stimulation functions while avoiding serious immune response in brain tissue. In this part, we will introduce the application of graphene in intracranial electrodes from two aspects: cortical electrodes based on graphene and implantable electrodes.

3.1 Cerebral cortex electrode based on graphene

cortical electrodes(ECoG electrodes)can record the integrated signals of multiple underlying neurons.At the same time,compared with implantable electrodes,Cortical electrodes cause less trauma to tissues and have been widely used in clinical judgment of neurological diseases(such as epilepsy).graphene has good transparency and excellent conductivity,which can be used to construct transparent ECoG electrodes.the combination of nerve electrode and optical imaging technology can link tissue observation with nerve signal detection,which is helpful to understand the structure and function of the brain more clearly and completely.in 2014,Park et al.prepared a transparentμECoG electrode array with 17 detection sites on a parylene-C substrate using four layers of Graphene(as shown in Figure 3A)^[78,79]。 Compared with the conventional metal electrode,the electrode showed higher transparency,and the blood vessels below the graphene site could be observed by fluorescence microscopy and optical coherence tomography(Fig.3B).they also used grapheneμECoG electrode arrays to record the response signals of the cortex of transgenic Thy1:ChR2 mice under different light intensities(Figure 3C).Therefore,graphene has a good application prospect in the field of transparent nerve electrodes.On this basis,They used the transparent graphene electrode as a stimulating electrode to regulate the neural electrical activity of the mouse cortex,and combined with calcium imaging technology in transgenic GCaMPf6 mice to map the spatial and temporal activity of neurons under electrical stimulation^[80]。 Kuzum et al.enhanced the conductivity of graphene by doping a single layer of graphene by exposing the graphene surface to concentrated nitric acid(70%),and further used the doped graphene to fabricate a transparentμECoG electrode array^[81]。 the doped graphene has the characteristics of large interface capacitance,small charge transfer resistance and the like,so that the signal-to-noise ratio of the nerve electrode is 5 to 6 times higher than that of a gold electrode with the same size。

the combination of transparent nerve electrode and photogenetic technology realizes the observation of the activity state of nerve cells at the electrode site under light stimulation.However,due to the photoelectric(Becquerel effect)and photothermal effect,light stimulation is often accompanied by light artifacts,which appear in the form of transients or oscillations in electrophysiological recordings,thus interfering with the measurement of electrophysiological signals.Thunemann et al.used the"bubbling method"transfer process and double-layer lithography technology based on polymethylglutarimide(PMGI)to prepare a crack-free and adhesive-free graphene transparent neural electrode array(as shown in Figure 3D)^[82]。 It was found that the electrode could eliminate the interference of Optical artifacts and directly record the electrophysiological activities of mouse cerebral cortex with high resolution and high signal-to-noise ratio(as shown In Figure 3e,3F).in addition,the study combines neural electrophysiological recording with optical techniques such as depth-resolved two-photon imaging and transgenic calcium imaging.optical observation of arteriolar dilatation and hemodynamic response under the electrode site induced by photogenetic stimulation was realized,thus extending the transparent graphene neural electrode to a universal platform suitable for a variety of different experimental environments^[82,83]。

Although the impedance of graphene can be reduced by chemical treatment such as plasma treatment and nitric acid doping,the effect of chemical treatment on impedance reduction is limited,generally only 2 to 10 times.this is mainly because the above treatment only reduces the Fermi level of graphene and increases its hole concentration.However,because the quantum capacitance of graphene is too small,that is,the density of Dirac point States is very low,the improvement of hole concentration by lowering the Fermi level is limited.to solve This problem,Lu et al.Developed a method to reduce the electrode impedance by electrodepositing platinum nanoparticles(PtNP)on graphene^[84]。 While maintaining optical transparency,the electrode breaks through the quantum capacitance limit and reduces the electrode impedance by 100 times.the graphene/PtNP microelectrode array can record nerve signals on the surface of the cerebral cortex and simultaneously carry out calcium ion imaging on cells at different depths.In addition to doping and modifying the graphene electrode to reduce its impedance,adjusting the morphology of graphene is also an effective way to reduce the impedance of the electrode.Lu et al.Prepared a flexible ECoG electrode array with three-dimensional porous graphene as the detection site on polyimide substrate by laser pyrolysis and nitric acid doping^[85]。 the electrode shows larger specific surface area and lower electrochemical impedance than the planar electrode.This electrode shows a high signal to noise ratio when recording low amplitude evoked potentials from the cerebral cortex。

the impedance of the electrode is not the only factor that determines the performance of the electrode,mainly because thermal noise associated with electrode impedance is not the only source of noise that affects neural recordings.electrical noise,biological noise,and other noise sources associated with the recording electrodes in the recording system greatly degrade the detection and classification of neural Electrical signals.Compared with the microelectrode array,the field effect transistor has the inherent property of signal amplification,which can effectively reduce the external noise and improve the sensitivity of the device^[86]。 Garrido et al.fabricated a graphene liquid gate transistor on a polyimide substrate using a single layer of graphene grown by chemical vapor deposition(as shown in Figure 3G).For the first time,the graphene liquid gate transistor was used to record the spontaneous slow wave activity(as shown in Figure 3H),the response signal triggered by the visual and auditory cortex and the synchronous discharge activity of epilepsy in rats under anesthesia^[87,88]。 they found that the signal-to-noise ratio of the graphene electrode was higher than that of the platinum black electrode in the low frequency range(as shown in Figure 3 I).Subsequently,They further studied the recording of cortical signals by graphene liquid-gate transistors,confirming that the graphene liquid-gate transistors can record neural signals in a wide electrophysiological bandwidth.Especially in the low frequency range(<0.1 Hz),the signal to noise ratio is higher,which is suitable for recording cortical diffusion inhibition signals and provides the possibility of recording low frequency signals with high fidelity and high spatial resolution^[89]。 in addition,In order to solve The problem of large contact resistance between metal as source and drain electrodes and graphene channels,and the problem of splitting delamination caused by the mismatch of Young's modulus between graphene and contact electrodes when flexible deformation occurs,our group has developed an all-carbon graphene transistor device.the transistor consists of a graphene channel and a source drain electrode of a graphene/carbon nanotube hybrid film^[90]。 Because the graphene in the channel and the graphene in the hybrid electrode are formed synchronously,a seamless connection structure is formed between the transistor channel and the source and drain electrodes.Moreover,compared with the traditional metal electrode,the mechanical properties of the graphene/carbon nanotube hybrid electrode and the graphene channel are more matched,so that the transconductance of the graphene does not change when the device is subjected to bending deformation.On this basis,we further prepared all carbon fiber nanoelectrodes by mechanical compression method^[91]。 Compare with a planar electrode,that electrode has a lower project area on the premise of keeping the same effective area,so that the device shows higher spatial resolution than a planar device when bee used For detecting rat epilepsy signals.for graphene field effect transistor electrodes,the rapid increase in the number of electrode sites leads to an increase in the space occupied by the connector,and the existing ECoG technology still does not provide the resolution to accurately map the high-level activity of the brain。

Schaefer et al.Used a multiplexing readout method in the graphene liquid gate transistor array,which made the increase of electrode sites more reasonable and reduced the complexity of device fabrication^[92]。 the distortion mechanism in graphene liquid gate transistors has also attracted the attention of researchers,which distorts the recording of signals with large amplitude or high frequency.Garcia-Cortabella et al.Have described and modeled in detail the harmonic distortion and non-ideal frequency response in graphene liquid gate transistors^[93]。 this precise description of the input-output relationship in a graphene liquid gate transistor allows the definition of a voltage and frequency dependent transfer function.Finally,This functional relationship is used to calibrate a variety of electrophysiological signals ranging from large amplitude low-frequency cortical spreading inhibition signals to low amplitude high-frequency action potentials。

as a practical tool for detecting brain activity,cortical electrodes can reveal the role of specific areas of the brain in body function.However,some specific information(such As precise movement)is presented by a small number of neurons in the form of a single potential peak.the electrophysiological signal of such a weak neural population can easily disappear in the cerebral cortex,so it can not be detected.This also limits the application of ECoG electrodes in the detection of weak electrophysiological signals。

3.2 Graphene-based implantable electrode

Compared with cortical nerve electrodes,implantable electrodes can form closer contact with neurons,thus detecting the electrophysiological signals of weak nerve clusters.By reducing the size of the electrode detection site,the electrical signal of a single neuron can be recorded.Due to its good biocompatibility and electrical properties,graphene can be Used as a surface coating of rigid nerve electrodes to reduce the immune response of brain tissue to nerve electrodes.for example,Bourrier et al.used a single layer of graphene film as a coating For a solar electrode^[94]。 the results showed that the proliferation of astrocytes and microglia around the graphene-coated nerve electrode was reduced after implantation,and the signal quality and durability of the nerve electrode were improved.Zhao et al.synthesized a seamless and full-coverage graphene coating on copper microwires by low-pressure chemical vapor deposition(as shown in Figure 4A),which effectively reduced the cytotoxicity of copper microwires(as shown in Figure 4B)^[95]。 the electrode can be used to record local field potentials and single cell action potentials In the rat brain(as shown in fig.4C)without image artifacts in an MRI scanner.However,the instability of the bond between the coating and the electrode may lead to coating exfoliation and electrode failure during long-term use.in order to avoid coating exfoliation,Garrett et al.Prepared soft reduced graphene oxide microfibers by wet spinning and used them as independent nerve stimulation and recording electrodes^[96]。 They used parylene-C to encapsulate reduced graphene oxide microfibers and laser ablated their front ends to form"brush"electrodes to improve the quality of neural interfaces.Later,Wang et al.prepared a low impedance graphene composite microelectrode by modifying Pt coating on graphene fibers(as shown in Fig.4D and 4E)^[97]。 in vivo studies have shown that electrodes implanted in the rat cortex can detect neuronal activity with high signal-to-noise ratios in areas as small as a single neuron(Figure 4 f)。

Similarly,graphene transistors can also be used as implantable neural electrodes for neural electrical signal recording.Our group has developed a dual-mode neural probe based on graphene transistors(as shown in Figure 4G)^[98]。 the electrode can be used both as a cortical electrode and as an implantable electrode(as shown in Fig.4h).This electrode was used to synchronously record the epileptic signals on the surface and inside of the cerebral cortex of rats.the results showed that the epileptic signal amplitude on the surface of the cerebral cortex was significantly greater than that inside the cortex,and the discharge signal on the surface of the cortex appeared earlier than that inside the cortex(as shown in Figure 4i),indicating that the discharge of epilepsy began at the cortical surface.in addition,Garrido et al.Also used dual-mode graphene transistors to detect the depth distribution of low-frequency voltage changes induced by diffuse inhibitory signals in the cerebral cortex^[89]。 It has been found that diffuse inhibitory signals in the cortex occur throughout the cortical depth,from long depolarization at the surface of the cortex to hyperpolarization at the depth.in recent years,although some progress has been made,the application of graphene in implantable nerve electrodes is still not common。

4 Conclusion and prospect

Graphene combines many characteristics suitable for neural electrode active materials, such as good compatibility, chemical stability, flexibility, optical transparency and high conductivity, which provides a new choice for the construction of bidirectional neural interface for synchronous detection and regulation. In this paper, the interaction between graphene and nerve cells, and the detection and stimulation of electrophysiological signals in nerve cells, brain slices and living brain using graphene microelectrode arrays and graphene transistors are reviewed. The results show that graphene has good compatibility with nerve cells, can promote the differentiation of neural stem cells into neurons, and can enhance the electrical signals of neural networks. Compared with the traditional rigid metal electrode, the flexible neural electrode prepared by transferring graphene to the flexible substrate can be better attached to the brain tissue, thus realizing the recording of neuronal electrical signals with high spatiotemporal resolution and high signal-to-noise ratio. In addition, the flexible transparent neural electrode based on graphene has both mechanical flexibility and transparency, which can simultaneously record optical imaging and electrophysiological signals, as well as photogenetic regulation of nerve cell activity under the electrode. Finally, as an implantable electrode, graphene can effectively reduce the immune response of brain tissue and improve the durability of nerve electrodes.

Although graphene provides new possibilities for the construction of ideal neural electrodes, there is still much room for the development of graphene-based neural electrodes: (1) In order to obtain higher signal-to-noise ratio and temporal and spatial resolution,To achieve high density and high throughput integration of flexible electrodes, appropriate surface treatment methods, pore structure design of graphene, multiplexing and wireless transmission technologies need to be further developed. (2) In order to realize the analysis of neural circuit mechanism, it is necessary to further develop the application potential of graphene materials in implantable neural electrode detection technology. (3) The new generation of neural electrodes need to provide multiple functions at the same time, including the recording and regulation of electrophysiological signals, the recognition of neurotransmitters and other neuro-related biomolecules, and the ability to control drug delivery. (4) Neural electrode technology needs to be combined with other brain regulation and detection technologies (such as optogenetic technology, magnetic resonance imaging, etc.), which is of great help to the study of brain mechanism. The development of multi-functional composite graphene electrodes with high sensitivity and stability and their application in basic neuroscience and the treatment of brain-related diseases involve chemistry, materials, medicine, engineering and other disciplines, which requires the joint efforts of researchers in many fields to achieve substantive innovative results and technologies.

References

Publishing order | Descend order by publishing year | Descend order by cited within

[1]	Chen R, Canales A, Anikeeva P. Nat. Rev. Mater., 2017, 2(2): 16093.

[2]	Moxon K A, Foffani G. Neuron, 2015, 86(1): 55.

[3]	Fattahi P, Yang G, Kim G, Abidian M R. Adv. Mater., 2014, 26(12): 1846.

[4]	Nicolelis M A L. Nat. Rev. Neurosci., 2003, 4(5): 417.

[5]	Stacey W C, Litt B. Nat. Clin. Pract. Neurol., 2008, 4(4): 190.

[6]	Du Z H, Lu Y, Wei P F, Deng C S, Li X J. Acta Phys. -Chim. Sin., 2020, 36(12): 2007004.

[7]	都展宏, 鲁艺, 蔚鹏飞, 邓春山, 李骁健. 物理化学学报, 2020, 36(12): 2007004.

[8]	Lacour S P, Courtine G, Guck J. Nat. Rev. Mater., 2016, 1(10): 16063.

[9]	Bouton C E, Shaikhouni A, Annetta N V, Bockbrader M A, Friedenberg D A, Nielson D M, Sharma G, Sederberg P B, Glenn B C, Mysiw W J, Morgan A G, Deogaonkar M, Rezai A R. Nature, 2016, 533(7602): 247.

[10]	Wallace G G, Moulton S E, Clark G M. Science, 2009, 324(5924): 185.

[11]	Cody P A, Eles J R, Lagenaur C F, Kozai T D Y, Cui X T. Biomaterials, 2018, 161: 117.

[12]	Rousche P J, Normann R A. J. Neurosci. Methods, 1998, 82(1): 1.

[13]	Xie F, Xi Y, Xu Q D, Liu J Q. Acta Phys. -Chim. Sin., 2020, 36: 2003014.

[14]	谢凡, 奚野, 徐庆达, 刘景全. 物理化学学报. 2020, 36: 2003014.

[15]

Jun J J, Steinmetz N A, Siegle J H, Denman D J, Bauza M, Barbarits B, Lee A K, Anastassiou C A, Andrei A, Aydın Ç, Barbic M, Blanche T J, Bonin V, Couto J, Dutta B, Gratiy S L, Gutnisky D A, Häusser M, Karsh B, Ledochowitsch P, Lopez C M, Mitelut C, Musa S, Okun M, Pachitariu M, Putzeys J, Rich P D, Rossant C, Sun W L, Svoboda K, Carandini M, Harris K D, Koch C, O’Keefe J, Harris T D. Nature, 2017, 551(7679): 232.

[16]	Csicsvari J, Henze D A, Jamieson B, Harris K D, Sirota A, BarthÓ P, Wise K D, Buzsáki G. J. Neurophysiol., 2003, 90(2): 1314.

[17]	Gray C M, Maldonado P E, Wilson M, McNaughton B. J. Neurosci. Methods, 1995, 63(1/2): 43.

[18]	He G W, Dong X F, Qi M. J. Funct. Mater., 2019, 50(12): 12026.

[19]	贺光伟, 董旭峰, 齐民. 功能材料, 2019, 50(12): 12026.

[20]	Brummer S B, Turner M J. IEEE Trans. Biomed. Eng., 1977, BME-24(1): 59.

[21]	McCreery D B, Agnew W F, Yuen T G H, Bullara L. IEEE Trans. Biomed. Eng., 1990, 37(10): 996.

[22]	Brunton E K, Winther-Jensen B, Wang C, Yan E B, Hagh Gooie S, Lowery A J, Rajan R. Front. Neurosci., 2015, 9: 265.

[23]	Cogan S F. Annu. Rev. Biomed. Eng., 2008, 10(1): 275.

[24]	Nick C, Quednau S, Sarwar R, Schlaak H F, Thielemann C. Microsyst. Technol., 2014, 20: 1849.

[25]	Im C, Seo J M. Biomed. Eng. Lett., 2016, 6(3): 104.

[26]	Polikov V S, Tresco P A, Reichert W M. J. Neurosci. Methods, 2005, 148(1): 1.

[27]	Salatino J W, Ludwig K A, Kozai T D Y, Purcell E K. Nat. Biomed. Eng., 2017, 1(11): 862.

[28]	Wang J F, Tian H H, Fang Y. Chin. J. Anal. Chem., 2019, 47(10): 1549.

[29]	王晋芬, 田慧慧, 方英. 分析化学, 2019, 47(10): 1549.

[30]	McConnell G C, Rees H D, Levey A I, Gutekunst C A, Gross R E, Bellamkonda R V. J. Neural Eng., 2009, 6(5): 056003.

[31]	Liu Y, Duan X J. Acta Phys. -Chim. Sin., 2020, 36: 2007066.

[32]	刘杨, 段小洁. 物理化学学报, 2020, 36: 2007066.

[33]	Gimsa J, Habel B, Schreiber U, van Rienen U, Strauss U, Gimsa U. J. Neurosci. Methods, 2005, 142(2): 251.

[34]	Rivnay J, Wang H L, Fenno L, Deisseroth K, Malliaras G G. Sci. Adv., 2017, 3(6): e1601649.

[35]	Xu K, Wang J F. Acta Phys. -Chim. Sin., 2020, 36: 2003050.

[36]	许可, 王晋芬. 物理化学学报, 2020, 36: 2003050.

[37]	Huang X, Yin Z Y, Wu S X, Qi X Y, He Q Y, Zhang Q C, Yan Q Y, Boey F, Zhang H. Small, 2011, 7(14): 1876.

[38]	Guan S L, Wang J F, Fang Y. Nano Today, 2019, 26: 13.

[39]	Mayorov A S, Gorbachev R V, Morozov S V, Britnell L, Jalil R, Ponomarenko L A, Blake P, Novoselov K S, Watanabe K, Taniguchi T, Geim A K. Nano Lett., 2011, 11(6): 2396.

[40]	Geim A K, Novoselov K S. Nat. Mater., 2007, 6(3): 183.

[41]	Nair R R, Blake P, Grigorenko A N, Novoselov K S, Booth T J, Stauber T, Peres N M R, Geim A K. Science, 2008, 320(5881): 1308.

[42]	Buzsáki G, Anastassiou C A, Koch C. Nat. Rev. Neurosci., 2012, 13(6): 407.

[43]	Yang K, Zhang S, Zhang G X, Sun X M, Lee S T, Liu Z. Nano Lett., 2010, 10(9): 3318.

[44]	Wen H Y, Dong C Y, Dong H Q, Shen A J, Xia W J, Cai X J, Song Y Y, Li X Q, Li Y Y, Shi D L. Small, 2012, 8(5): 760.

[45]	Yang K, Wan J M, Zhang S, Zhang Y J, Lee S T, Liu Z. ACS Nano, 2011, 5(1): 516.

[46]	Shan C S, Yang H F, Han D X, Zhang Q X, Ivaska A, Niu L. Langmuir, 2009, 25(20): 12030.

[47]	Gollavelli G, Ling Y C. Biomaterials, 2012, 33(8): 2532.

[48]	Li N, Zhang X M, Song Q, Su R G, Zhang Q, Kong T, Liu L W, Jin G, Tang M L, Cheng G S. Biomaterials, 2011, 32(35): 9374.

[49]	Agarwal S, Zhou X Z, Ye F, He Q Y, Chen G C K, Soo J, Boey F, Zhang H, Chen P. Langmuir, 2010, 26(4): 2244.

[50]	Kanayama I, Miyaji H, Takita H, Nishida E, Tsuji M, Fugetsu B, Sun L, Inoue K, Ibara A, Akasaka T, Sugaya T, Kawanami M. Int. J. Nanomed., 2014, 9(1): 3363.

[51]	Park S Y, Park J, Sim S H, Sung M G, Kim K S, Hong B H, Hong S. Adv. Mater., 2011, 23: H263.

[52]	Zuidema J M, Hyzinski-García M C, van Vlasselaer K, Zaccor N W, Plopper G E, Mongin A A, Gilbert R J. Biomaterials, 2014, 35(5): 1439.

[53]	Wang Y, Lee W C, Manga K K, Ang P K, Lu J, Liu Y P, Lim C T, Loh K P. Adv. Mater., 2012, 24(31): 4285.

[54]	Kim J, Park S, Kim Y J, Jeon C S, Lim K T, Seonwoo H, Cho S P, Chung T D, Choung P H, Choung Y H, Hong B H, Chung J H. J. Biomed. Nanotechnol., 2015, 11(11): 2024.

[55]	Bendali A, Hess L H, Seifert M, Forster V, Stephan A F, Garrido J A, Picaud S. Adv. Healthc. Mater., 2013, 2(7): 929.

[56]	Fabbro A, Scaini D, LeÓn V, Vázquez E, Cellot G, Privitera G, Lombardi L, Torrisi F, Tomarchio F, Bonaccorso F, Bosi S, Ferrari A C, Ballerini L, Prato M. ACS Nano, 2016, 10(1): 615.

[57]	Veliev F, Briançon-Marjollet A, Bouchiat V, Delacour C. Biomaterials, 2016, 86: 33.

[58]	Pasca S P. Nature, 2018, 553(7689): 437.

[59]	Ma Q Q, Yang L Y, Jiang Z Y, Song Q, Xiao M, Zhang D, Ma X, Wen T Q, Cheng G S. ACS Appl. Mater. Interfaces, 2016, 8(50): 34227.

[60]	Li N, Zhang Q, Gao S, Song Q, Huang R, Wang L, Liu L W, Dai J W, Tang M L, Cheng G S. Sci. Rep., 2013, 3: 1604.

[61]	Song Q, Jiang Z Y, Li Ning, Liu P, Liu L W, Tang M L, Cheng G S. Biomaterials, 2014, 35: 6930.

[62]	Fang Q J, Zhang Y H, Chen X B, Li H, Cheng L Y, Zhu W J, Zhang Z, Tang M L, Liu W, Wang H, Wang T, Shen T, Chai R J. Front. Bioeng. Biotechnol., 2020, 7: 436.

[63]	Guo W B, Qiu J C, Liu J Q, Liu H. Sci. Rep., 2017, 7: 5678.

[64]	Sakai K, Teshima T F, Nakashima H, Ueno Y. Nanoscale, 2019, 11(28): 13249.

[65]	Qian Y, Zhao X T, Han Q X, Chen W, Li H, Yuan W E. Nat. Commun., 2018, 9: 323.

[66]	Tang M L, Song Q, Li N, Jiang Z Y, Huang R, Cheng G S. Biomaterials, 2013, 34(27): 6402.

[67]	Pampaloni N P, Lottner M, Giugliano M, Matruglio A, D’Amico F, Prato M, Garrido J A, Ballerini L, Scaini D. Nat. Nanotechnol., 2018, 13(8): 755.

[68]	Cohen-Karni T, Qing Q, Li Q, Fang Y, Lieber C M. Nano Lett., 2010, 10(3): 1098.

[69]	Hess L H, Jansen M, Maybeck V, Hauf M V, Seifert M, Stutzmann M, Sharp I D, Offenhäusser A, Garrido J A. Adv. Mater., 2011, 23(43): 5045.

[70]	Blaschke B M, Lottner M, Drieschner S, Calia A B, Stoiber K, Rousseau L, Lissourges G, Garrido J A. 2D Mater., 2016, 3(2): 025007.

[71]	Cheng Z G, Hou J F, Zhou Q Y, Li T Y, Li H B, Yang L, Jiang K L, Wang C, Li Y C, Fang Y. Nano Lett., 2013, 13(6): 2902.

[72]	Kireev D, Zadorozhnyi I, Qiu T, Sarik D, Brings F, Wu T R, Seyock S, Maybeck V, Lottner M, Blaschke B M, Garrido J, Xie X M, Vitusevich S, Wolfrum B, Offenhäusser A. IEEE Trans. Nanotechnol., 2017, 16(1): 140.

[73]	Kireev D, Seyock S, Ernst M, Maybeck V, Wolfrum B, Offenhäusser A. Biosensors, 2017, 7(1): 1.

[74]	Rastogi S K, Bliley J, Shiwarski D J, Raghavan G, Feinberg A W, Cohen-Karni T. Cel. Mol. Bioeng., 2018, 11(5): 407.

[75]	Chen C H, Lin C T, Chen J J, Hsu W L, Chang Y C, Yeh S R, Li L J, Yao D J. 2011 16th International Solid-State Sensors, Actuators and Microsystems Conference. IEEE, 2011: 1883.

[76]	Chen C H, Lin C T, Hsu W L, Chang Y C, Yeh S R, Li L J, Yao D J. Nanomed.: Nanotechnol. Biol. Med., 2013, 9(5): 600.

[77]	Du X W, Wu L, Cheng J, Huang S L, Cai Q, Jin Q H, Zhao J L. J. Biol. Phys., 2015, 41(4): 339.

[78]	Kireev D, Seyock S, Lewen J, Maybeck V, Wolfrum B, Offenhäusser A. Adv. Healthcare Mater., 2017, 6(12): 1601433.

[79]	Jeong D W, Kim G H, Kim N Y, Lee Z, Jung S D, Lee J O. RSC Adv., 2017, 7(6): 3273.

[80]	Cheng J, Wu L, Du X W, Jin Q H, Zhao J L, Xu Y S. J. Microelectromech. Syst., 2014, 23(6): 1311.

[81]	Veliev F, Han Z, Kalita D, Briançon-Marjollet A, Bouchiat V, Delacour C. Front. Neurosci., 2017, 11: 466.

[82]	Kireev D, Brambach M, Seyock S, Maybeck V, Fu W Y, Wolfrum B, Offenhäusser A. Sci. Rep., 2017, 7: 6658.

[83]	Schaefer N, Garcia C R, Calia A B, Mavredakis N, Illa X, Masvidal C E, Cruz J, Corro E, Rodriguez L, Prats A E, Bousquet J, Martinez A J, Perez M A P, Hebert C, Villa R, Jimenez D, Guimera B A, Garrido J A. Carbon, 2020, 161: 647.

[84]	Park D W, Schendel A A, Mikael S, Brodnick S K, Richner T J, Ness J P, Hayat M R, Atry F, Frye S T, Pashaie R, Thongpang S, Ma Z Q, Williams J C. Nat. Commun., 2014, 5: 5258.

[85]	Park D W, Brodnick S K, Ness J P, Atry F, Krugner-Higby L, Sandberg A, Mikael S, Richner T J, Novello J, Kim H, Baek D H, Bong J, Frye S T, Thongpang S, Swanson K I, Lake W, Pashaie R, Williams J C, Ma Z Q. Nat. Protoc., 2016, 11(11): 2201.

[86]	Park D W, Ness J P, Brodnick S K, Esquibel C, Novello J, Atry F, Baek D H, Kim H, Bong J, Swanson K I, Suminski A J, Otto K J, Pashaie R, Williams J C, Ma Z Q. ACS Nano, 2018, 12(1): 148.

[87]	Kuzum D, Takano H, Shim E, Reed J C, Juul H, Richardson A G, de Vries J, Bink H, Dichter M A, Lucas T H, Coulter D A, Cubukcu E, Litt B. Nat. Commun., 2014, 5: 5259.

[88]	Thunemann M, Lu Y C, Liu X, Kılıç K, Desjardins M, Vandenberghe M, Sadegh S, Saisan P A, Cheng Q, Weldy K L, Lyu H M, Djurovic S, Andreassen O A, Dale A M, Devor A, Kuzum D. Nat. Commun., 2018, 9: 2035.

[89]	Liu X, Lu Y C, Iseri E, Shi Y H, Kuzum D. Front. Neurosci., 2018, 12: 132.

[90]	Lu Y C, Liu X, Hattori R, Ren C, Zhang X W, Komiyama T, Kuzum D. Adv. Funct. Mater., 2018, 28(31): 1800002.

[91]	Lu Y C, Lyu H M, Richardson A G, Lucas T H, Kuzum D. Sci. Rep., 2016, 6: 33526.

[92]	Kireev D, Offenhäusser A. 2D Mater., 2018, 5(4): 042004.

[93]	Blaschke B M, Tort-Colet N, Guimerà-Brunet A, Weinert J, Rousseau L, Heimann A, Drieschner S, Kempski O, Villa R, Sanchez-Vives M V, Garrido J A. 2D Mater., 2017, 4(2): 025040.

[94]

HÉbert C, Masvidal-Codina E, Suarez-Perez A, Calia A B, Piret G, Garcia-Cortadella R, Illa X, del Corro Garcia E, de la Cruz Sanchez J M, Casals D V, Prats-Alfonso E, Bousquet J, Godignon P, Yvert B, Villa R, Sanchez-Vives M V, Guimerà-Brunet A, Garrido J A. Adv. Funct. Mater., 2018, 28(12): 1703976.

[95]

Masvidal-Codina E, Illa X, Dasilva M, Calia A B, Dragojević T, Vidal-Rosas E E, Prats-Alfonso E, Martínez-Aguilar J, de la Cruz J M, Garcia-Cortadella R, Godignon P, Rius G, Camassa A, del Corro E, Bousquet J, HÉbert C, Durduran T, Villa R, Sanchez-Vives M V, Garrido J A, Guimerà-Brunet A. Nat. Mater., 2019, 18(3): 280.

[96]	Shi E Z, Li H B, Yang L, Hou J F, Li Y C, Li L, Cao A Y, Fang Y. Adv. Mater., 2015, 27(4): 682.

[97]	Yang L, Zhao Y, Xu W J, Shi E Z, Wei W J, Li X M, Cao A Y, Cao Y P, Fang Y. Nano Lett., 2017, 17(1): 71.

[98]	Schaefer N, Garcia-Cortadella R, Martínez-Aguilar J, Schwesig G, Illa X, Moya Lara A, Santiago S, HÉbert C, Guirado G, Villa R, Sirota A, Guimerà-Brunet A, Garrido J A. 2D Mater., 2020, 7(2): 025046.

[99]	Garcia-Cortadella R, Masvidal-Codina E, de la Cruz J M, Schäfer N, Schwesig G, Jeschke C, Martinez-Aguilar J, Sanchez-Vives M V, Villa R, Illa X, Sirota A, Guimerà A, Garrido J A. Small, 2020, 16(16): 1906640.

[100]

Bourrier A, Shkorbatova P, Bonizzato M, Rey E, Barraud Q, Courtine G, Othmen R, Reita V, Bouchiat V, Delacour C. Adv. Healthcare Mater., 2019, 8(18): 1970075.

[101]

Zhao S Y, Liu X J, Xu Z, Ren H Y, Deng B, Tang M, Lu L L, Fu X F, Peng H L, Liu Z F, Duan X J. Nano Lett., 2016, 16(12): 7731.

[102]

Apollo N V, Maturana M I, Tong W, Nayagam D A X, Shivdasani M N, Foroughi J, Wallace G G, Prawer S, Ibbotson M R, Garrett D J. Adv. Funct. Mater., 2015, 25(23): 3551.