Preparation, types, and applications of one- and two-dimensional nanochannels and their transport properties for water and ions

2.1 Preparation of 2D materials

2D materials, including elemental and compound, can be divided into two types according to the way they are bonded between the layers of 2D materials: layered materials and non-layered materials. The controlled preparation of 2D materials and the exploration of their physical properties and new functional applications on this basis not only have important scientific research value, but also promote their wide application. At present, there are three main types of controlled preparation methods for 2D materials: micromechanical exfoliation, chemical vapor deposition (CVD), and liquid exfoliation.

2.1.3 Liquid-phase stripping method

In the liquid-phase exfoliation method, a nanosheet dispersion of 2D materials is prepared by overcoming the interlayer adhesion of 2D materials through a certain external force in water or organic solvents [12]. The liquid-phase exfoliation method mainly includes the following methods: ion intercalation method, ion selection method, and ultrasonic-assisted exfoliation method.

Ion interpolation uses ions in the liquid phase to be inserted into the layers of bulk materials to increase the spacing between crystal layers, thereby weakening the interlayer interactions, and finally completely exfoliates them with the assistance of applied external fields (such as shear force, ultrasonic, or thermal expansion) to form a 2D material nanosheet dispersion, such as the preparation of graphene oxide (GO) by the Hummers method [13] and the electrochemical intercalation of molybdenum sulfide (MoS₂) by n-butyllithium intercalation or quaternary ammonium molecules [14]. Ion intercalation exfoliation is often accompanied by the creation of vacancies or functionalization, and the resulting 2D nanosheets are uniformly dispersed in water or organic solvents. For MoS₂ exfoliation, macromolecular quaternary ammonium salt molecules (such as tetraheptylammonium bromide) were first inserted into the bulk MoS₂ layer by electrochemical methods, and then, a 2H-phase MoS₂ nanosheet dispersion was prepared by exfoliation, with a concentration of up to 10 mg ml⁻¹.

In the ion displacement method, some layered bulk materials carry ions between the layers, which can be displaced in the liquid phase and assisted agitation to obtain a dispersion of 2D material nanosheets, such as peeled vermiculite [15], layered double hydroxides. The ion replacement method mainly selects cations with large, hydrated ion radius (such as Li ions) to replace the intrinsic cations between the layers, and at the same time, more water molecules enter, increase the layer spacing, and complete the peeling. Taking vermiculite stripping as an example, the 2D vermiculite nanosheet dispersion was obtained by two-step cation substitution, Li ions were selected to replace the Mg ions contained in the interlayer of the block, and the 2D vermiculite nanosheet dispersion was obtained through ultrasonic treatment. Or select anions with small binding force (such as Cl ions), replace the anions with strong binding force in the bulk material, and select suitable solvents to complete the peeling. Before and after peeling, the surface structure and charge distribution of the 2D material sheet remain unchanged, because they can be evenly and stably dispersed in the solvent due to the electrostatic repulsion of each other.

In the ultrasound-assisted exfoliation method, high-energy ultrasound is used to weaken the adhesion between the sheets. One method is to select a solvent with suitable surface energy to prevent the aggregation and precipitation of the peeled 2D material nanosheets, and directly ultrasonic peeling to obtain the dispersion of the 2D material nanosheets. Another method is to use surfactants, which are homogeneously dispersed in water with nanosheets of 2D material. For example, sodium cholate was used as the surfactant to directly ultrasound the bulk 2D materials, WS₂, MoTe₂, MoSe₂, NbSe₂, TaSe₂, and boron nitride (BN), to prepare the corresponding 2D nanosheet dispersions [16].

The preparation of 2D materials by liquid-phase exfoliation method has the characteristics of a wide sourcing of raw materials, simple process, low cost, and can obtain a large amount of 2D materials, which is suitable for industrial production. The exfoliated 2D materials are dispersed in water or organic solvent in the form of liquid phase, and the concentration is adjustable, the yield is high, and different three-dimensional macroscopic structures can be constructed by liquid-phase treatment. Saikam et al. [17] synthesized exfoliated graphite (EG) using liquid-phase exfoliation by changing the heating source or by changing the intercalation compounds, and the general synthesis process and some potential applications are depicted in Figure 1.

Figure 1

Schematic diagram of synthesis and applications of EG [17].

2.2 1D nanochannels

Nanomaterials are classified into three different types according to their size and morphology: zero-dimensional nanoparticles, 1D nanotubes, and 2D nanosheets. The cross-sectional length and width of 1D nanochannels are less than 100 nm, and they mainly include 1D nanopores and nanotubes [34]. In terms of preparation methods, 1D nanochannels can be divided into 1D nanochannels formed by bottom-up assembly of monomers, and 1D nanochannels formed by physical or chemical etching of materials, such top-down preparation methods include polymer nanochannels prepared by ion beam etching technology and solid-state nanopores formed by electron beam exposure [35].

2.3 Preparation of 2D nanochannels

2D nanochannels are 2D structural channels formed in materials, which are characterized by having a height or thickness in the nanoscale, while the length and width can be larger, forming a 2D confined space. These channels enable the selective transport of molecules or ions and therefore have a wide range of application potential in areas such as sensors, energy storage, and separation technologies. For example, rapid, selective, and sensitive detection of contaminants can be achieved by integrating ordered 2D COFs into nanofluid channels [46]. The self-assembled films of 2D materials have a typical layered structure, which is characterized by a parallel arrangement of highly oriented nanosheets. The 2D material film’s uniqueness is that it has many 2D nanochannels inside, and the channel size is comparable to that of the 2D material nanosheets. These 2D nanochannels are connected to each other, resulting in a complete nanopathway that runs through the entire film. Methods for assembling 2D layered membranes mainly include a variety of combination methods, such as suction filtration, coating relying on substrate smoothness and solution surface tension, and LBL self-assembly with alternating deposition of different 2D materials on the substrate.

3.1 Desalination

Rapid social progress, especially industrialization and urbanization, has exacerbated the problem of water pollution. Population growth and climate change are increasing the demand for fresh water, and there is an urgent need to address freshwater scarcity. Considering that 97% of the planet’s water resources are seawater, desalination offers a potential solution that has a broader prospect than the development of groundwater. At present, in the industrial field, distillation and reverse osmosis technology are mainly used for seawater desalination. Distillation (including multi-stage flash and multi-effect distillation) is an energy-intensive, environmentally significant, large-scale, and costly process for hot seawater desalination. In contrast, reverse osmosis uses external force or osmotic pressure to pass seawater through a specific polymer membrane, effectively separating ions to produce fresh water, which is both energy-saving and environmentally friendly. However, it is difficult for traditional polymer membranes to balance water permeability and ion selectivity. Low-dimensional nanomaterials have opened up a new research direction for seawater desalination technology due to their excellent water permeability and ionic selectivity after special treatment and have great commercial potential.

Corry [55] used classical molecular dynamics (MD) simulations to show that the hydrophobic and smooth (6, 6) CNT can completely intercept ions and achieve 100% desalination rate under the action of 200 MPa hydrostatic pressure. The desalination rate of (7, 7) CNT, which is larger, can also reach 97%, and the water permeability is significantly higher than that of (6, 6) CNT at the same pressure. At present, the industry still encounters significant challenges in the large-scale synthesis of defect-free, high-density, and high-orientation CNTs. With the advancement of sustainable development strategies, energy costs and environmental pollution have become the focus of attention. The high energy consumption of seawater desalination may be a better solution by removing ions from seawater. By applying an external voltage, ions are temporarily stored in an electrical bilayer formed by porous carbon nanomaterials, which can significantly reduce the energy consumption of the desalination process compared to reverse osmosis [56]. Some researchers have also found that the interaction between alkali metal cations and large bonds will block ion channels and seriously reduce the permeability of water, resulting in uncertainty about the application prospect of CNT in water desalination [57]. Using classical MD, Cohen-Tanugi et al. [58] revealed the mechanism of desalination of single-layer graphene through pore size, as shown in Figure 2. Their research discovered that nanopores in a single graphene layer can filter NaCl salts from water. By adjusting the pore size, they allowed water to pass through while blocking ions, thus achieving desalination. At present, the industry has not been able to synthesize high-orientation, high-density, and defect-free nanotubes on a large scale. Although experiments can produce nanoscale pore sizes by ion or electron bombardment, precise control of hole shape and size is still very difficult.

Figure 2

Schematic diagram of MD simulation of water desalination [58].

3.2 Ion separation

Biological ion channels play an extremely important role in cell transport. The process and selectivity mechanisms by which ions pass through membrane protein channels are complex, especially for biological protein channels, which are affected by factors such as steric hindrance, Coulomb interactions, charged amino acid arrays, and ion dehydration. Biological ion channels are essential for cell transport. The process and selectivity mechanism of ions passing through membrane protein channels are particularly complex, especially in biological protein channels, which are affected by multiple factors such as steric hindrance, Coulomb interaction, charged amino acid arrangement, and ion dehydration. Although biochannel proteins have excellent ionic selectivity, their industrial applications are significantly limited due to their fragile mechanical properties and bioenvironment-dependent activity. The synthetic nanochannels are narrow and hydrophobic, with excellent mechanical properties and large specific surface area. The chemical stability, mechanical strength, and electrical conductivity in the structure give it the potential for a wide range of applications in ion separation devices. Researchers have conducted a lot of theoretical and experimental studies on the application of CNT in ion separation, explored the ion selectivity mechanism of CNT, and analyzed the influence of steric hindrance, Coulomb coupling, and other factors on ion transport and separation.

The charge on the surface of CNTs has a significant effect on the selectivity and transport of charged ions. Aluru et al. [59] demonstrated that asymmetrically functionalized (16, 16) CNTs pass significantly higher chloride ions than potassium ions through MD simulations. Majumder et al. [60] placed negatively charged functional groups at the nozzle of the CNT and found that this could significantly improve the cation transport efficiency. Gong et al. [61] showed that (9, 9) the inner surface of CNTs can be modified with different carbonyl functional groups to achieve the separation of Na⁺/K⁺ ions. Siria et al. [62] fabricated a hierarchical nanofluidic device consisting of a single boron nitride nanotube (BNNT) that connects two reservoirs across an impermeable solid-state membrane to study diffusion and osmosis, as shown in Figure 3.

Figure 3

Nanotube with smooth inner walls used as an ion transport channel [62]: (a) sketch of the transmembrane BNNT for nanofluidic measurements and (b) its experimental realization, imaged by transmission electron microscopy.

With the development of nanotechnology, the area and quality of graphene prepared experimentally are getting larger and higher, and the method of generating pores on a porous matrix has been developed, which makes porous graphene have a good application prospect in the field of ion separation. Thanks to the development of industrial technology, layered graphene nanochannels with precise layer spacing have been successfully prepared experimentally, which has expanded the application of nano-flow devices in the field of separation.

3.3 Energy harvesting

Since the beginning of the Industrial Revolution, fossil fuels (mainly coal, oil, or natural gas) have been the main source of energy for humanity. However, burning fossil fuels is a serious pollution to the environment, and burning fossil fuels produces a large amount of greenhouse gasses, causing global warming, melting glaciers, rising sea levels, abnormal climate, and frequent El Niño phenomena around the world. Therefore, access to sustainable, abundant, and affordable energy is one of the major challenges facing modern society. Potentially available renewable energy sources include sunlight, wind, tidal, and geothermal. However, so far, none of these energy sources have been able to replace fossil fuels, mainly because of the limited efficiency of generating and storing electricity, medium availability, unsustainability, and high cost. Scientists and engineers are also exploring tidal energy, wave friction power generation, and blue energy. Using the difference in salinity between seawater and river water (salinity gradient) to generate clean energy, known as blue energy or osmotic energy, is a promising, challenging, accessible, clean renewable energy source.

At present, pressure-retarded osmosis and reverse electrodialysis (RED) are two possible methods for extracting and converting osmotic energy (Gibbs free energy). The differential pressure permeation method uses the osmotic pressure generated by the different salinities on both sides of the semi-permeable membrane to drive water molecules from the freshwater tank to the seawater tank, and the flowing water can drive the connected turbine generator to generate electricity. However, the average power (1 W m⁻²) of the separation membrane produced by this method in actual industry is much lower than the target power (5 W m⁻²) [63], so the development of this project has been suspended. Electrodialysis (RED) refers to the direct generation of electrical energy by a salinity gradient-driven ion through a selective nanofluidic device of cation or anion. In models using silica nanochannel configurations, this power can even reach ∼8 W m⁻². The Dutch company REDS take is currently experimenting with this approach to generate electricity on a large scale, and the membranes are now achieving conversion efficiencies in the range of a few watts per square meter, which is close to commercial targets.

Feng et al. [64] developed a monolayer porous MoS₂ membrane with a pore size of 10 nm, which can optimistically estimate a power density of 106 W m⁻² in the case of massive parallelization. Siria et al. [65] discussed current technologies for converting blue energy. Specifically, the membrane serves as an ionic current generator, as illustrated in Figure 4. The generated electric power is collected by connecting electrodes in the reservoir to an external load resistance. However, unlike seawater desalination or ion separation, electroosmotic energy-enrichment nanochannels do not require strict ion separation, and the pore size of nanochannels can reach tens of nanometers without precise pore size regulation [62,65], which greatly increases the possibility of their industrial applications.

Figure 4

Diffusion–osmotic process for osmotic energy conversion [65].

Although highly conductive electrodes and ion-exchange membranes have been developed for energy recovery, practical applications still face several challenges. For example, the interaction between the nanopore ion-exchange membrane and the salt solution may affect the storage of ions during subsequent desalination, reducing the efficiency of energy recovery. Therefore, optimizing the interfacial properties of ion-exchange membranes can help improve the efficiency of energy recovery.

3.4 Single-molecule detection

In 1953, Coulter invented the Cooler counter and was the first to develop the resistance pulse method based on the ion transmission principle to achieve the analysis and detection of single particles. When a single particle enters a micron-to-millimeter-sized pore under electric field migration, the current decreases due to the blockage of the ionic current, and when it moves out of the pore, the current recovers, resulting in a resistor-pulse signal in the I–T curve. By analyzing the frequency, peak width, and amplitude of the ionic currents during perforation, information on particle size distribution, concentration, shape, and surface charge can be obtained [66,67]. In recent years, solid-state nanopores have become a versatile alternative to bionanopores due to their clearly adjustable geometry and pore size, high mechanical strength, ease of modification, and compatibility with various devices and technologies. DNA is composed of repeating units of nucleotides, and only one nucleotide can reside in ultra-thin graphene pores at a time, so sequencing can be performed by observing the translocation of successive nucleotides through the nanopores [68].

The research team of Jiang et al. [69,70,71] summarizes recent progress in artificial solid-state nanochannel-based sensing systems for detecting various types of biomarkers. Artificial solid-state nanochannels with probes on the outer surface not only contribute to the in situ detection of large analytes but also provide promising opportunities for ultra-high sensitivity detection with a clear mechanism. They developed a set of intricately designed supersandwich DNA probes and fabricated wettability-controllable nanochannels to explore the mechanism of nanochannels with regional wettability modification. Their research shows that this single set of probes can effectively facilitate both reliable on-site qualitative detection and ultra-sensitive quantitative analysis, all while minimizing error propagation. They consider that designing solid-state nanochannels with functionalized outer surfaces and achieving quantum-confined superfluid would effectively overcome the challenge of universal single-molecule detection and ultra-high-precision spatiotemporal resolution. These pioneering nanochannels will expand the scope of biosensor applications and facilitate the exploration of mass transfer mechanisms.

In 2021, Gao et al. [72] reported a protocol to explicitly divide the regional and functional elements at the outer surface and inner wall of solid-state nanochannels (SSNs) in order to achieve the highly specific ionic output in complex matrices. This will contribute to the fundamental understanding of substance transport in SSNs and provide high specificity and sensitivity in SSN-based analyses. In other works, in order to study the signal mechanism, Wu et al. [73,74] used electroneutral peptide nucleic acid and negatively charged DNA to quantitatively analyze the individual contributions of the space charge and steric effects of charged targets on the outer surface to the ionic signal in order to demonstrate the signal mechanism and used the rolling circle amplification (RCA) process to synthesize DNA amplicons to enhance the ionic signal through the space charge effect. Their research shows that the ionic signal increases with the local surface potential per nanochannel and is further enhanced by decreasing the porosity density of nanochannels.

In 2022, Si et al. [75] reported a method using polydopamine (PDA) to modify the highly charged surface in nanoconfinements in order to fix the electrostatic repulsion in nanoconfines, which limits ion current rectification (ICR) manipulation. Their research found that the charged properties of PDA change the ICR effect of the nanochannel due to property changes in the ion-depletion zone from PDA modification.

Meller’s team [76] improved the DNA capture rate of SiN nanopores by establishing a transmembrane ion concentration gradient, and a 20-fold salt gradient could detect picomolar concentrations of DNA at high throughput. Thakur and Movileanu [77] discovered that specific and nonspecific interactions in a heterogeneous solution can be distinguished using a protein bait-containing biological nanopore in mammalian serum, as shown in Figure 5. Wu developed an analytical strategy-based nanochannel system modified with RCA amplicons on the outer surface, which provides a highly sensitive and specific sensing platform for identifiable molecules using nucleic acids.

Figure 5

Single-molecule protein detection at single-tethered receptor resolution [77].

4.1 Graphene nanochannels

In the 2D confined channel, scholars have prepared graphene nanochannels based on the material properties of graphene, and their dimensional accuracy can be controlled at 0.34 nm channel, and the channel wall has the characteristics of atomic-level smoothness, which has extremely high water transport properties. In GO films, oxygen-containing functional groups on GO sheets are used as supports to form graphene nanochannels with atomic-level smoothness in the layers of unoxidized sheets stacked on top of each other. MD studies have shown that the mechanism of the rapid passage of water molecules through GO films is not driven by traditional differential pressure, but by the combined effect of capillary pressure and evaporation to drive water molecules to enter graphene nanochannels [80].

Experimental studies have shown that GO films are size-selective in terms of ion screening. The graphene nanochannels in the film are only a few layers of aqueous molecules perpendicular to the wall and are therefore capable of precise size selectivity for ions in aqueous solutions as well as other solute molecules. MD structural simulations verify this phenomenon – allowing water and small ions to pass through quickly and large ions to tissue to pass through. And the experimentally measured ion diffusivity is several thousand times higher than the value estimated using the classical concentration diffusion theory. Through simulation studies, researchers have discovered the phenomenon of “ion sponge effect” [47], i.e., the interaction between graphene and ions causes ions to aggregate in the nanochannels, forming a higher concentration gradient, which further promotes the ion diffusion process. Shi et al. [81,82] via the drop-casting method to prepare freestanding reduced GO (rGO) from an rGO suspension in order to observe Na₂Cl and Na₃Cl as 2D Na–Cl crystals on reduced GO membranes (rGOM), and they used the MD simulations method to incorporating cation–p interactions to classic all-atoms force fields in order to show that Na⁺ has a clear enrichment on a carbon-based p electron-rich surface in NaCl solutions. Their research found that Na⁺ is the p electron-rich structures in aqueous salt solutions, and the cation–π interactions between the ions and the aromatic rings in the graphitic surfaces would promote the ion adsorption on the surface to form crystals and help to make 2D Na–Cl crystals of stoichiometries.

Li et al. [83] found that there are strong electrostatic, van der Waals, and hydrogen bond interactions between water/ions and oxygen-containing groups in GO nanosheets, which largely hinder mass transport of species. Therefore, the rapid water/ion transport in GO membranes mainly occurs in the non-oxidizing region adjacent to the GO nanosheets. Gu et al. [84] showed that the electrostatic interaction of ions with the surface of fluorinated graphene (F-GRA) results in ideal ion repulsion on the surface of F-GRA and ultra-low water friction mediated by its hydrophobicity, resulting in good water permeability. Esfandiar et al. [85] prepared graphene electrodes with a thickness of nearly 100 nm by mechanical exfoliation, constructed a single sub-nanometer channel with a bilayer of graphene as a spacer, and found that the conductivity of the chloride solution in the sub-nanometer channel was much lower than that of the bulk solution, and the water molecules tended to face the outward direction of the –OH functional group in the hydrated layer of Cl– ions, and the surface of graphene was found to promote the polarization of water molecules and strongly attract water molecules, resulting in a strong force between the two, making Cl– ion transport characteristics are greatly reduced.

GO flakes have both oxidized and non-oxidized bands. According to slip theory, water molecules exhibit frictionless, ultra-fast flow in the non-oxidized zone of the GO layer, resulting in a very fast water permeation rate. The oxidized band, containing abundant oxygen functional groups, undergoes strong hydrogen bonding interactions with water molecules, hindering their rapid slip [86]. The transport mechanism of ions is mainly due to steric hindrance and cation–π interactions on GO nanosheets. Researcher [87] shows that by controlling the interlayer spacing between GO flakes, the size of nanochannels can be adjusted, allowing for precise separation of different ions and molecules. Additionally, GO films with nanochannel sizes larger than the ionic hydration radius exhibit good ion repulsion performance. The primary component of cation–π interaction is electrostatic, inducing a cross-linking effect that helps control interlayer spacing in GOMs. The cation–π interaction was first proposed by Sun et al. [88], who prepared GO films using a drip method. Metal cations (including K⁺, Na⁺, Ca²⁺, Li⁺, Mg²⁺) can reduce the interlayer spacing of GO films to 0.1 nm [89]. The interaction between cations and π is very strong in non-covalent interactions. Chen et al. [90] used the drop-casting method to prepare GO GOM in order to separate ions from bulk solution or for sieving ions of a specific size range from a mixed salt solution. Combining experimental observations and theoretical calculation, their research found that cations (K⁺, Na⁺, Ca²⁺, Li⁺ and Mg²⁺) themselves can determine and fix the interlayer spacing of GOMs at sizes as small as a nanometer and the variable range of this spacing can be controlled to within one ångström.

Due to the unique spatial arrangement of water molecules in the confined space and the solid–liquid contact area-to-volume ratio much higher than that of macroscopic fluids, the interaction between water and nanochannels is extremely important in phase transition, and the phase transition temperature of water is affected by the structure of the channel pipeline and is different from that of body water, which also has high application value in practical engineering. Experimental studies have shown that when the separation between solid surfaces approaches the atomic scale, fluid properties change dramatically. In particular, the drying transition occurs on the surface of water molecules. This is due to the action of strong hydrogen bonds between water molecules that cause the liquid molecules to retreat. The non-polar surface forms different layers that separate the bulk phase from the surface, so that the hydrogen bonding network of water molecules in the confined space exhibits multiple phase transition states when the hydrogen bonding network is broken [91]. Simulation studies have found that water molecules can freely enter the nanochannels formed by two graphene sheets with a gap of only about 0.35 nm and spontaneously form a 2D crystalline solid structure, i.e., a single layer of square ice, which is a new form of water at room temperature. MD simulations have revealed that the van der Waals force interaction between graphene sheets exerts a confined pressure of up to the order of GPa on the confined water, which changes the hydrogen bonding network of the water from a 3D to a 2D plane, and the orderly arrangement of water molecules results in a 2D square crystal structure [92]. Raju et al. [93] studied the phase transitions of water in 2D graphene nanochannels at scales of 5.5, 8.5, and 11.5 Å and observed that water forms an ordered hexagonal phase at low densities and an ordered square phase at high densities. Through the analysis of structural properties by radial distribution function and transverse density profle, it is observed that the double-layer hexagonal ice has the characteristics of the first-order phase transition, while the double-layer (three-layer) square ice structure has no phase transition, which is a continuous structural change without a clear transition temperature.

4.2 Micromechanical nature of separation pressure and boundary slip

The separation pressure caused by the solid–liquid interface at the nanoscale [94–96] provides an additional driving force for fluid transport, which is essentially the embodiment of various intermolecular forces between the solid–liquid interfaces. Intermolecular forces at the nanoscale are key to explaining the rapid transport of nanofluids. The nanoscale has an ultra-high specific surface area; there is a strong surface interface effect; incompressible Newtonian fluid, wall no slip, and other conditions may no longer be applicable; and the microscopic transport process cannot be described by classical theory.

The flow behavior of non-viscous liquids without kinetic energy loss is superfluidity, which is typically characterized by the fact that the substance is not viscous, there is no friction with the contact surface, and it can be transported in the channel without resistance. Superfluidity was first discovered in liquid helium by Kapitsa [97] and Kim and Chan [98], and a theoretical model was established. When the fluid flows on the solid wall, the fluid atoms at the solid–liquid interface will be limited to the equilibrium position of the solid lattice structure due to the force from the solid surface, the liquid–liquid interaction force inside the fluid will cause the fluid atoms on the surface to deviate from the equilibrium position, and the fluid atoms will gradually deviate from the equilibrium position of the solid surface when the liquid–liquid interaction is greater than the solid–liquid interaction, showing the slip phenomenon. Existing studies have shown that the graphene nanochannels in GO films can allow water molecules to pass through quickly, and nano-water droplets can quickly diffuse on the surface of graphene with regular morphology; thanks to the long boundary slip length (60–100 nm) of the graphene channel wall and the hydrogen bonding network of water in the confined space, the coupling of wall action and surface charge in the nanochannel is the key to affect material transport and ion screening. In smooth-walled nano-sized channels, the transition of water molecular configuration enhances the flow, and the frontage effect due to channel roughness is enhanced. Defects and functional groups, for example, can eliminate the flow-enhancing effect [99]. The water molecules flowing between the GO sheets are hydrophilic groups in the oxidized region, which hinders the rapid transport of water molecules, while the water molecules in the non-oxidized graphene sheets form a 2D network structure between them, and the fluidity is significantly enhanced [80]. Crystals with unconventional stoichiometries hold promise for the design of novel transistors, magnetic devices, and catalyzers, Xia et al. [100] reported a highyield synthesis of 2D abnormal Na₂Cl crystals base on applying a a negative potential on a rGOM which not only promoted the ion adsorption on the surface to form crystals but also helped to stabilize 2D Na₂Cl with an excess of sodium. Their research expand crystals with unconventional stoichiometries research interest both in the laboratory and in applications. Their research found that with the cation–π interaction that can be directly strengthened in a crystallization process, the adsorption capacity and stabilization of the crystals will be greatly enhanced. Wu et al. [101] from China University of Petroleum comprehensively analyzed the MD simulation results and experimental data from 53 studies and found that in nanochannels with different wettability and pore size, the viscosity of water is inversely proportional to the contact angle, and the contact angle in the hydrophilic pipeline is small, the viscosity is large, and the flow velocity is smaller than that of body water. Hydrophobic pipes have large contact angles, low viscosity, and flow velocities that are even seven orders of magnitude larger than bulk water. Chen et al. [102] linked the layered structure formed by water molecules confined to the interlayer voids of graphene with their rapid transport characteristics, and they believed that this special layered structure was the main reason for the rapid transport of water molecules between layers. Falk et al. [103] from France studied the relationship between friction coefficient and radius in water flow between graphene sheets, inside armchair CNTs, and outside these CNTs. They discovered that the flow velocity was enhanced in confined spaces. Additionally, the narrower the layer spacing, the more pronounced the enhancement, as illustrated in Figure 6. Jiao and Xu [104] found that when the temperature was lower than 315 K and the interlayer distance was 0.65 nm, a single layer of water appeared between the graphene sheets, and due to the reduction of wall friction, the water diffused in the graphene layer was discontinuous and collective, which was conducive to the rapid transport of molecules. Suk and Aluru [105] used MD simulations to study the flow of water in graphene nanopores and found that the pressure in the channel showed a linear downward trend while the rate of the fluid along the axial direction barely changed, confirming the heterogeneous structural properties of water in confined spaces, as well as the corresponding changes in density and viscosity.

Figure 6

Effects of confinement and curvature on friction in nanochannels [103].

Although scholars have conducted a lot of theoretical simulation research in the field of water transport, it is difficult to obtain a unified self-consistent conclusion due to the differences in simulation methods and parameters. However, due to the shielding of CNTs to the internal fluid, the experimental data for the study of fluid transport in CNTs are extremely limited, and it is difficult to systematically explain the physical mechanism of high-flux transport characteristics.

4.3 Size effect of nanochannel-confined mass transfer

Due to the influence of solid–liquid interaction and spatial confinement effect, the density distribution of the fluid along the radial direction of the nanochannel is not uniform, but the density near the channel wall is significantly higher than that in the center of the channel. Similar to the density distribution, the shear viscosity distribution within the nanochannel may also be non-monotonic. Solid–liquid interactions increase the local shear viscosity of fluids close to the wall, resulting in slow flow. The overlap and mutual interference of the two boundary layers close to the wall in a confined space further lead to a complex change in the viscosity of the fluid with the size of the channel. Based on its special size effect, nanochannels make it possible to separate substances, and the separation of molecules or ions of different sizes can be realized using the difference in the migration velocity of molecules of different sizes in the channel, and the limiting size effect of nanochannels makes it easier for substances to interact with the wall surface after passing through the nanochannels.

Studies have shown that when the size of nanopores or nanochannels is large at the micrometer scale, the transport characteristics of transmembrane ionic molecules are mainly related to their mobility in bulk solution. However, when the size of nanopores or nanochannels decreases sharply to the nanoscale, the transport characteristics of ionic molecules are changed by the action of the pore wall or channel wall. In addition, the proportion of transmembrane ionic molecules interacting increases with the decrease in pore wall or channel wall size, which means that the interaction of ionic molecules through nanoscale pores or channels cannot be ignored. Ions are present in solution in the form of hydrated ions, and when faced with smaller channels, more energy is required to remove the water molecules from the hydrated shell, which determines whether the ions can pass through or not. The use of spatial bits to block the passage of larger substances is an important feature to reflect the selectivity of nanochannels. Generally speaking, the hydrated ion structure formed by the combination of ions and water molecules is divided into the first water shell and the second water shell, and the number of water molecules contained in it is defined as the first coordination number and the second coordination number. When the pore size of the electrically neutral nanochannel is close to or smaller than the diameter of the ionic water shell, a huge energy barrier is formed for the ions that will enter it, preventing the ions from entering. A series of studies by Sun et al. [106] showed that for neutral sub-nanochannel ions, the dehydration of the first water shell is the origin of the ion transmembrane energy barrier. However, the recent separation of Mg⁺/Li⁺ ions in salt solution by Liu et al. [107] showed that the dehydration of the ionic second water shell is also the origin of the ion transmembrane energy barrier, and the dehydration of the second water shell should not be ignored. Gao et al. [108] also confirmed that the dehydration of the second water shell of the ion is the origin of the ion transmembrane energy barrier. Dai et al. [109] developed a class of GOMs from GO suspensions using vacuum filtration without drying treatment in order to have high water permeance when the ion rejection rate is at a satisfactory level and stable performance. Their research found that the ions can control and fix the interlayer spacings in the GOMs, and the distorted hydration structures of the multivalent ions due to the strong hydrated cation–p interaction, resulting in the higher rejection rates of undistorted hydrated ions in solutions through the GOMs. Wei et al. [110] prepared rGOM by the amino-hydrothermal method and the modified Hummers’ method from GO suspensions in order to separate of radioactive ⁹⁰Sr from other radioactive ions efficiently. Their research found that the distort energy barrier of Co²⁺ is significantly higher than that of Sr²⁺, making it difficult for hydrated Co²⁺ to distort and enter the internal space between GO sheets from the bulk solution due to the large energy barrier of dehydration.

Radha et al. [111] fabricated narrow, smooth capillaries using van der Waals assembly, as shown in Figure 7. Their findings indicate that when the channel height exceeds six atomic layers (2.38 nm), water molecules behave as predicted by classical fluid dynamics. However, when the height is below this threshold, the flow rate significantly surpasses predictions of traditional models. Zhao et al. [112] combined non-equilibrium molecular dynamics simulations and first principles density functional theory calculations to study the molecular flux of Lennard-Jones-benzene fluid in nanoslit channels and found a slip length oscillation caused by molecular size effects. Borg et al. [113] investigated the structural changes of water molecules within CNTs with diameters ranging from 0.81 to 4 nm. With the gradual increase of the size of the nanochannels, the structure of water molecules transitions from a single-chain structure to an ordered ring structure and finally changes to a structure like that of bulk water.

Figure 7

Schematic diagram of the nanoslit channel structure [111].

In the coming decade, the lack of clean water is a formidable challenge because of rapid population growth, extended droughts, and fast growing demands. Nanofiltration membranes (NFMs) are widely used in water purification due to its low energy cost and simple operational process. Generally, most of polymeric NFMs have advantages of flexibility, simple preparation process, and relatively low cost, but also face some problems, such as poor chemical resistance, limited lifetime, and membrane fouling. Thus, researchers proposed CNT-based NFMs, but it is still far away from practical application due to relatively high cost of CNTs, time-consuming, and complex process of obtaining high density of vertically aligned CNTs, and difficulties for achieving large-scale production. Therefore, 2D graphene materials have entered the field of vision of researchers, and graphene can form highly ordered films with 2D nanochannels between two graphene sheets by the facile filtration-assisted assembly method. But Nair et al. [80] prepared GOM using Hummer’s method in order to investigate molecular permeation through submicrometer-thick GOM. Their research found that submicrometer-thick membranes made from GO can be completely impermeable to liquids, vapors, and gases, but these membranes allow unimpeded permeation of water attribute to a low-friction flow of a monolayer of water through 2D capillaries formed by closely spaced graphene sheets. It seems that NFMs cannot be prepared by dried graphene films. Han et al. [114] used chemically converted graphene (CCG) to fabricate ultra-thin graphene NFMs with 2D nanochannels in order to be applied them for water purification. Their research shows that 34 mg of CCG is sufficient for making a square meter of NFM, indicating that this new generation graphene-based nanofiltration technology would be resource-saving and cost-effective.

Unlike the resistance existing in graphene sheets, GO can be easily dispersed in various solvents due to its unique surface chemical composition and amphiphilic nature. In particular, GO has excellent stability in water-soluble dispersion and has been realized in batch production, so GO has been used as the raw material for the production of graphene-based 2D membrane materials in many research works. The study of GOMs is still an emerging field, and research to understand the atomic structure, transport mechanism, and separation mechanism is in progress. Any change in dimensions of the nanochannels would result in dramatically different rates of transport of fluids through the GOMs. Aher et al. [115] used the drop-casting method to prepare the GOM in order to measure the permeation and rejection performance of the GOMs with isopropyl alcohol and water as solvents. Their research suggests that water adsorption on the GO surface is a critical factor that controls the flux and sieving effect of the GOM. Huang et al. [116] used a novel nanostrand-channeled strategy to prepare nanostrand-channeled GO (NSC-GO) ultrafiltration membranes to meet the growing worldwide demand for pressure-driven ultrafiltration membranes. Their research found that the separation performance of NSC-GO ultrafiltration membranes exhibits an abnormal dependence on applied pressure due to the reversible deformation of the nanochannels.

The rapid development of micro-nanotechnology has brought new challenges to mechanical experiments, and at microscales, the influence of size effects, multi-field (force, electricity, magnetism, heat) coupling, and non-classical forces (van der Waals force, capillary force, electrostatic force, Casimir force, etc.) have become factors that need to be considered, and many macroscopic solid mechanics experimental methods are difficult to apply. Therefore, it is of great scientific significance to carry out the research on the analysis method of micro/nanomechanical testing technology and multi-scale correlation. The internal structure of concrete provides a migration channel for medium transport, and the macro-meso-microscale structure characteristics of concrete determine the chloride ion transport behavior. De Vera et al. [117] experimentally investigated the effect of water saturation on chloride diffusion in concrete and found that when the water saturation was reduced from 95 to 54%, the diffusion coefficient of chloride ions decreased by about two orders of magnitude. Homan et al. [118] confirmed the coupling effect between moisture transport and chloride permeation in concrete by comparing the effects of moisture transport on chloride permeation in partially saturated and fully saturated concrete. Liu et al. [119] studied the transport characteristics of multiple groups of separators in the pores of the C–S–H gel under saturation through MD simulation and found that the calcium/silicon ratio had a significant effect on the adsorption of chloride ions on the surface of the C–S–H gel, and the adsorption was strongest when the calcium/silicon ratio was 1.2.

5.1 Theoretical model

Molecular simulation is a method of simulating macroscopic matter and system by constructing a microscopic system containing a large number of particles in a computer, solving the equation of motion of particles with the help of computer computing microsystem using the principles of quantum mechanics, and finally using the principle of statistical physics to calculate the various physical and chemical properties of macroscopic system materials at different time and space scales.

MD is a simulation method based on Newton’s equations of motion. By deducing the position of atoms and optimizing the interaction force between atoms, the optimal results of the system are obtained. Statistical mechanics is then used to obtain the macroscopic properties of the system, such as binding constants and molecular conformation. The MD method can not only study the static thermodynamic properties of macromolecular systems, but also study the dynamic mechanical properties, obtain various physical and chemical properties, and obtain data that are difficult to obtain under experimental conditions. Using computer-supported theories, the transition states and reaction mechanisms of chemical reactions can be studied, and molecular and material design can be carried out more efficiently, thereby shortening the material development cycle and reducing research costs. However, the details of the atomic-scale simulation of the MD method cannot be described, and it cannot be used to elucidate the chemical reaction process. It is more suitable for analyzing molecular models with fixed molecular species and bond formation.

Monte Carlo (MC) is a research method that transforms practical problems into probabilistic problems, which can also be called random simulations. The core idea of this method is random distribution, which generates a random molecular model from random numbers and makes the particles in the model move irregularly through a computer. Count whether the energy in the newly generated molecular model decreases, if it decreases, the model proves to be valid, and the next iteration is carried out until the final model meets the preset conditions. Compared with the MD method, the MC method does not need to calculate the intermolecular forces, so the simulation is simpler and faster. However, because the motion process of particles cannot be expressed, the dynamic information of the system cannot be obtained, and it is not suitable for large polymer systems.

The quantum MD approach is a density functional theory based on the Koohn–Sham equation and a de novo computational MD approach around density functional theory. There are three approximations, namely, the non-relativistic approximation, the single-electron approximation, and the adiabatic approximation. Scholars usually use quantum MD methods to achieve the accurate solution of the potential energy surface and the determination of the standing point structure of the potential energy surface, and for larger systems, only the local minimum point and transition state on the search reaction path are generally calculated. Purely based on quantum mechanical methods, dynamical simulations are generally only suitable for the dynamic behavior of very small systems on very short time scales, and their high computational costs limit their scope of application, because they are often very computationally large and are not suitable for studying systems with more than 1000 atoms.

Molecular mechanics are mainly based on classical mechanics, also known as force field methods. According to the Born–Oppenheimer approximation, molecular mechanical methods ignore the motion of electrons and express the resulting energy (e.g., bond energy) as an empirical analytical form of the coordinates of the atomic nucleus. These different forms of analytic functions are force field functions, which are at the heart of molecular mechanics. Mechanical parameters can be obtained by experimental measurements or fitting based on quantum mechanical calculations. Since the 80 s of the last century, with the advancement of computer technology, people have developed a variety of force fields with resonant potential energy expressions that contain more details, with similar function compositions and specific expressions. However, because the electron motion is ignored, it is not possible to simulate properties based on differences in electron distribution and phenomena based on changes in electron distribution, so many specialized force fields have also been developed, such as bond level potentials for semiconductors and hydrocarbons [122], ReaxFF force fields for simulating chemical reactions [123], etc. The calculation accuracy of molecular mechanics is largely determined by the selected force field form and related parameters, so it is necessary to carefully select and calibrate the force field suitable for the research system before all molecular mechanics simulations.

Deep learning is hierarchical feature learning based on neural network algorithms, as a technology to realize machine learning, which makes artificial intelligence produce revolutionary breakthroughs. From the structural point of view, in the network framework of deep learning, the input layer is used to receive data, the output layer is used to generate the final result, and there are zero or more hidden layers between the input layer and the output layer to learn the nonlinear mapping relationship between the input and output, and deep learning is usually composed of a plurality of hidden layers, which is composed of a deep nested network structure, which can activate a neuron multiple times and automatically discover the features required for the learning task, so the deep learning model is superior to the shallow machine learning model and the traditional data analysis method, and the deep learning can use its automatic feature learning ability to directly complete the modeling task on the high-dimensional original input data into better handle large-scale, noisy, and unstructured data. At present, deep learning has shown excellent performance in the fields of image, language, computer vision, etc., and many scholars have applied it to MD simulation.

First-principles calculation is a calculation method that directly solves the Schrödinger equation through some basic physical constants, such as electron mass, electron charge, Planck constant, and speed of light, and then obtains the structure and physicochemical properties of materials. There are two approximations, the Born–Oppenheimer approximation and the Hartree–Fock approximation. Since there is no need to provide experimental parameters for calculation, this method can be exempted from the limitation of experimental conditions, and then, the material properties can be explored more deeply from a microscopic perspective, and some phenomena that are difficult to observe experimentally can be obtained. First-principles can be used to calculate the dynamic behavior of atoms at the microscopic scale that cannot be accurately observed experimentally, as well as the complex electronic structure of materials, such as strong correlation and bonding. With the rapid development of computers, first-principles calculations have been widely used in MD simulations and have become an indispensable technical means in simulation research.

5.2 Software environment

In the development of MD computer simulation, the main work has been divided into two aspects. On the one hand, there is a strict theoretical foundation and a suitable mathematical calculation model, and on the other hand, the calculation principle of the mathematical model is written through the computer machine language to realize multithreaded computing on the computer.

Materials Studio molecular simulation software is an early commercial software that conveniently combines quantum mechanics, molecular mechanics, mesoscale simulation, analytical instrument simulation, and statistical correlation into one easy-to-use modeling environment. It allows people to create and study models of molecular structures, taking advantage of powerful graphical capabilities to present people’s results. And the software has corresponding service guidance, has a humanized visual interface, has a high degree of professionalism, and does not require researchers to have high computer ability requirements. However, due to the different systems for which software is developed, the fields of production are different, and they are not comprehensive. Moreover, due to the commercial secrecy mechanism, users cannot understand the core principle, which leads to the limitation of dealing with the problem and, at the same time, has a high cost of use.

Lammps, which stands for large-scale atomic/molecular massively parallel simulator, is a powerful MD calculation program developed by Steve et al. of the Sandia National Laboratory in the United States. In the early days, Fortran was written in Fortran, but with successive versions and enhancements, Lammps is now written in C++ code and is updated regularly every month. Lammps does not have a visual interface, and all operations are operated by manually writing IN input files to operate relevant settings and solving problems through computer background processing and calculation. The IN-script code can be used to initialize the simulation system, define the initial state atoms, set the specific simulation parameters, and start the simulation. However, Lammps does not provide intelligent data analysis and data plotting tools for MD simulations, and the maintenance is poor.

VMD software is a powerful molecular visualization and molecular simulation software, which uses OpenGL to provide high-quality 3D molecular graphics for displaying, animating, and processing large atomic simulation systems, and can be used in conjunction with a variety of mainstream MD simulation software. Written in C++ and developed for the interactive graphical display of molecular systems, especially biopolymers such as proteins or nucleic acids, VMD supports the display of dynamic data, can interact with a separate MD simulation application to provide a graphical user interface and visualization console for the simulation program, can display molecules in a user-selectable rendering style, and produces high-quality hardcopy images.

GROMACS is a molecular simulation package based on the GROMOS software package, which was developed to simulate biomolecules in solution. As one of the most widely used open-source MD simulation software, GROMACS has high computational efficiency, wide application range, friendly operation interface, and simple operation. His input file must be readable ASCII text, annotated, and provide an interface to the standard file formats used in the fields of biology, physics, and chemistry. The output file is generic, allowing the user to write their own analysis program or use the analysis tools provided by the gromacs software. However, gromacs software is not compatible with other MD simulation software.

CP2K is an open-source electronic structure and MD software package for massively parallel and linear-scaled electronic structure methods and state-of-the-art de novo computational MD simulations. It has a good quantum mechanical approach and can calculate many different properties. But the use of CP2K requires a long and complex input file containing all the required parts, keywords, and atomic types and their sits for the system to be calculated.

The user can obtain the required structure and properties by analyzing the.xyz trajectory file and the.out output file generated by the simulation.

Table 1 lists the commonly used software for MD simulations, describes their performances, and notes their limitations.

6.1 Research status

With the development of nanotechnology, functional nanomaterials have a wide range of application prospects in high-tech fields such as information, biology, energy, environment, and aerospace. CNTs, graphene, and other nanofluidic devices have great application potential in the fields of water desalination, ion separation, energy harvesting, detection, and inspection due to their good mechanical properties, extremely thin thickness, chemical stability, and environmental friendliness. The confinement effect reduces the entropy of the fluid, making its transport properties in the nanochannel different from that of the bulk phase. Factors such as steric hindrance, Coulomb interaction, and pore geometry have an important impact on the transport efficiency of water in nanochannels. In addition, the ions form a stable hydrated shell in an aqueous solution, and the radius of the hydrated ions is greater than that of the water molecules. Choosing the right size of nanochannels can completely block the passage of ions, allowing only water molecules to pass through. Studying the osmotic transport of ions at the atomic scale can help develop effective methods for water filtration and desalination. Biological proteins have good selectivity for ions, but their selection process is complex, and it is difficult to distinguish the role of various factors in ion selection. However, inorganic nanochannels are relatively easy to control the variables, and the roles of ion dehydration, Coulombic interaction, and functional groups in ion selection can be explored. The study of the mechanism of liquid-phase transport and separation has important scientific value and guiding significance for the industrial application of nanofluidic devices in the future.

6.2 Trends

In view of the shortcomings and challenges of the current research, the future research trends of the confinement effect of nanochannel ion transport can be summarized as follows:

1. The behavior of water within the nanoscale channel is significantly different from that at the macroscale, where the motion-governing equations of classical fluid mechanics can no longer account for the flow behavior of water. With the continuous advancement of material preparation technology, we are now able to precisely control the eigenvalue of nanochannels. For example, graphene 2D channels have atomically smooth channel walls, and channel heights can be reduced to the angstrom level. These angstrom-scale height channels can be directly compared to experiments by simulation. Studying the transport of ions in nanochannel systems has long been of interest to the academic community due to their importance in understanding the properties of environmental, biological, and chemical processes.

2. Several successful studies have shown the potential to construct nanochannels on ion-exchange membranes to achieve selective separation of specific ions. However, how to successfully fabricate selective nanochannels on ion-exchange membranes and the mechanism of their selectivity still need to be further explored. In addition, long-term membrane stability is critical under real-world operating conditions, requiring high chemical, mechanical, and thermal durability, and therefore requires extensive research to achieve stability in membrane structure and performance. Although the membranes studied in these studies have specific ion selectivity at a laboratory scale, upgrading them to a low-cost industrial scale remains a significant challenge. Overall, artificial nano-sized ion channels provide a promising solution for efficient target ion separation and energy conversion.

3. With the deepening of nanofluidics research, the manipulation of fluid transport behavior is gradually developing in the direction of more miniaturization. In nanochannels, the regulation of fluid transport behavior is mainly based on the electrodynamic principle for the regulation of molecular or ion transport behavior inside the fluid. The size of the nanochannel is similar to the thickness of the electric double layer of the channel, which makes it possible to regulate the transport behavior of molecules according to their charged state. According to the structure and function relationship of nanochannels in biofilms, the biomimetic design of functionalized artificial nanochannels has great application potential in regulating the behavior of nanoscale fluids and designing functionalized nanofluid devices.

4. The size of nanochannel materials is small, and it is difficult, costly, and blind to directly control and realize functional design and modification experiments. Therefore, it is necessary to theoretically prove the feasibility of the experiment before conducting it. With the development of computer technology and the enrichment of theoretical models, MD simulation methods have gradually developed into one of the important methods for experimental design and prediction. The simulation results can provide direct guidance for the experiment of materials, improve the success rate of the experiment, and reduce the blindness of the experiment, and the development of MD simulation has great potential in the study of nanochannel experiments.

6.3 Conclusion and outlook

In this study, we introduce the progress of industrial applications of nanochannels, such as desalination, ion selection, energy enrichment, and detection, and explore the phase behavior of confined water in nanochannels, the micromechanical mechanism of confined mass transfer, and material modification in combination with mainstream theoretical models and MD simulation software. Among them, the influence of the nanochannel confinement effect on the confined water structure, shear viscosity, self-diffusion, and hydrogen bonding behavior is revealed, and the micromechanical nature of the confined water flow boundary slip in the nanochannel is clarified based on atomic-scale structure and morphological analysis, which greatly expands the content of the mechanical properties of the nanoscale ion transport confinement conditions. Therefore, future research will seek more breakthroughs in the following channel applications to promote the mechanical properties of nanochannel ion transport confinement conditions to practical applications.

Graphene nanochannels in GO films have the ability to transport water molecules quickly, but are strictly limited by nanochannel size and film stability. How to stably control the size of graphene nanochannels is the key to realize high-flux graphene-based filtration membranes, which provides new ideas for the design of seawater desalination membranes and other separation and mass transfer processes based on size selectivity.

1. Nanochannels are a key component of nanofluid devices, and the function of nanochannels determines the application of nanofluid devices. Therefore, how to design and fabricate functionalized nanofluidic devices is one of the key issues restricting the development of nanofluidics. How to biomimetic the structure of artificial nanochannels based on solid materials, according to the characteristics of biological channel structure, by regulating the interaction between the channel and the transported material (electrostatic interaction, van der Waals interaction, hydrogen bond interaction, etc.), to achieve intelligent regulation of the transported substances in the channel.

2. Water under extreme confinement has attracted extensive attention because its structural phase and dynamic behavior are significantly different from those of bulk water. Many researchers have found that the enhancement of the water transport rate is mainly due to its low-friction flow on graphene. The smooth surface properties of graphene and the interface effect between water and graphene are important factors that lead to low friction. Therefore, the study of the phase state and dynamic behavior of confined water structures is helpful to reveal the mechanism of the high-speed water flow through the graphene nanocapillary channels observed in the experiments.

With the advancement of ion transport mechanism research and biomimetic design of solid-state nanochannels, the industrial application of nanochannels has become an unavoidable topic. Future development should be more diversified and integrated to provide more possibilities for scientific research and technological innovation.