Advance in additive manufacturing of 2D materials at the atomic and close-to-atomic scale

Abstract

Atomic and close-to-atomic scale manufacturing (ACSM) has emerged as promising technologies in the manufacturing paradigm. Among various materials, 2D materials have garnered significant attention for ACSM due to their atomic-scale characteristics and physical properties. While chemical vapor deposition (CVD) can be employed to produce high-quality 2D materials, achieving patterning often relies on photolithography techniques, which limit scalability and introduce impurities. To address these challenges, this article serves as a review by focusing on exploring atomic-scale additive manufacturing methods for 2D materials. Several potential techniques are reviewed, including site-selective CVD, area-selective atomic layer deposition, electrodeposition, laser-assisted synthesis, print methods, and atomic layer-aligned stacking. The applications of atomic-scale additive manufacturing in various fields, such as electronics, biosensing, and nanoelectromechanical systems, are discussed. Finally, the future prospects of atomic-scale additive manufacturing for 2D materials based on existing research are delved into.

Overview of atomic and close-to-atomic scale manufacturing

Based on historical development, manufacturing technologies can be divided into three phases: Craft-based manufacturing by hand with millimeters or sub-millimeters precision; Precision-controllable manufacturing by machinery with micrometers and even nanometers precision; atomic and close-to-atomic scale manufacturing (ACSM) with atomic precision^1,2,3,4,5. These advances happened in different periods, but technologies developed exist in parallel, thereby representing three paradigms of manufacturing advancement. ACSM, as the fundamental technology of Manufacturing III, can add, remove and transform materials at the atomic scale^{6,7,8,9,10,11}. In subtractive manufacturing, precise techniques like ultra-precision machining, nanolithography, high energy beam-based direct writing, and tip-based machining are employed to achieve nanometric precision in dimensions and surface finish^12,13. This approach involves working with multiple materials, including metals, semiconductors, and 2D materials¹⁴. On the other hand, additive manufacturing (AM) focuses on layer-by-layer material deposition to create intricate structures. Techniques like atomic layer deposition, self-assembly with proteins, peptides, and deoxyribonucleic acid (DNA) are utilized, with the goal of achieving atomic precision in single-layer deposition or constructing complex bio-structures through controlled chemical reactions. The materials utilized in AM range from metals¹⁵, metal oxides, semiconductors, and 2D materials to biomacromolecules. Lastly, transformative manufacturing involves the manipulation of single atoms to compose complex structures. This method relies on altering atomic bonds in single crystalline materials or 2D materials, allowing for precise atomic-scale manipulations. It is obvious that ACSM is still confined itself to limited materials. Ideally, the target material for ACSM should possess an atomic structure and exhibit the potential for processing at the atomic scale. In this context, 2D materials have garnered significant interest.

2D materials are composed of single or few layers of atoms, held together by strong covalent bonds within each layer and weak van der Waals forces between the layers. This bonding structure allows for the isolation of materials down to a single atomic layer. Since the mono-atomic layer graphene was obtained by exfoliating with Scotch tape in 2004¹⁶, other families of 2D materials, such as Transition-metal dichalcogenides (TMDs) and Hexagonal boron nitride (h-BN), have also gained attention for their exceptional properties. Graphene has supreme mechanical stiffness, strength and elasticity, and very high electrical and thermal conductivity¹⁷. Molybdenum disulfide (MoS₂) is a representative TMDs material, and monolayer 2H-MoS₂ has a direct bandgap of 1.9 eV, which has outstanding electronic and photonic properties¹⁸. Hexagonal boron nitride (h-BN) has an atomically smooth surface without dangling bonds and charge traps, and it is an ideal substrate for other 2D materials¹⁹. Many studies have proved the superiority of the performance of 2D materials in the fields of electronics, biomedicine, sensors and other potential applications. However, the bottleneck that hinders the mass application of 2D materials is the synthesis and integration of 2D materials into functional devices.

The manufacturing challenges in achieving ACSM for 2D materials can be primarily divided into two aspects: the synthesis of high-quality atomic layer films of 2D materials and their subsequent patterning & integration. Common synthetic methods for 2D materials include chemical vapor deposition (CVD) and atomic layer deposition (ALD)^20,21. The quality of the synthesis of 2D materials has made great progress, and the latest methods can reach the wafer level with CVD and ALD on metal, sapphire, and silicon substrate^22,23,24,25. This enables close to single atomic layer addition to a semiconductor substrate with potential towards functional device fabrication. Conventional patterning methods for 2D materials, such as photolithography and etching, provide the necessary precision and a mature manufacturing process. However, these methods entail a complex process and have limitations in preventing the potential introduction of impurities.

In contrast, direct additive manufacturing of 2D materials at the atomic scale offers an alternative approach that can circumvent the need for photolithography, leading to improved production efficiency, process simplification, and avoiding impurities. Meaningful atomic-scale additive manufacturing needs to meet two simultaneous conditions: first, controlling the synthesis position of 2D materials on the plane, achieving a certain dimensional resolution, facilitating photolithography-free patterns, and simplifying the traditional process flow. Second, it is imperative to control the thickness of a single layer or a few layers of 2D materials, ensuring good layer resolution. This control is crucial for eliciting the various properties of 2D materials and meeting the requirements of various applications. The ultimate objective is to realize high-quality 3D structures of 2D materials with controllable position and thickness.

The latest advances in additive manufacturing of 2D materials towards ACSM are systematically analyzed in this article. It is aimed to explore the potential of additive manufacturing techniques for area-selective development of 2D materials while ensuring precise control over the number of layers. Firstly, typical manufacturing methods that show promise for achieving atomic-scale additive manufacturing of 2D materials are discussed. These methods include site-selective chemical vapor deposition (CVD), area-selective atomic layer deposition (ALD), electrodeposition, laser-assisted synthesis, print method and atomic layer aligned stacking. Then, the applications that necessitate atomic-scale additive manufacturing of 2D materials are presented. These applications exploit the various properties and functionalities of 2D materials, benefiting from the precise control over layer thickness and site achievable through additive manufacturing techniques. The methods and applications are shown in Fig. 1. Finally, the future prospects of atomic-scale additive manufacturing for developing 2D materials are delved into, showing its potential impact and the primary research challenges that need to be addressed.

Fig. 1: Atomic scale additive manufacturing of 2D materials and applications.

Source materials image – Reproduced from ref. 46 with permission from American Chemical Society, Copyright 2018. Reproduced from ref. 51 with permission from American Chemical Society, Copyright 2009. Reproduced from ref. 63 with permission from American Chemical Society, Copyright 2020. Reproduced from ref. 59 with permission from John Wiley and Sons, Copyright 2017. Reproduced from ref. 77 with permission from American Chemical Society, Copyright 2022. Reproduced from ref. 76 with permission from American Chemical Society, Copyright 2020. Reproduced from ref. 89 with permission from Royal Society of Chemistry, Copyright 2012. Reproduced from ref. 102 with permission from John Wiley and Sons, Copyright 2022. Reproduced from ref. 118 with permission from American Chemical Society, Copyright 2022. Reproduced from ref. 119 with permission from American Chemical Society, Copyright 2011. Reproduced from ref. 143 with permission from IOP Publishing, Copyright 2014. Reproduced from ref. 148 with permission from IOP Publishing, Copyright 2014. Reproduced from ref. 132 with permission from Springer Nature, Copyright 2017.

Additive manufacturing methods of 2D materials

Various bottom-up manufacturing methods for atomic-level two-dimensional materials are systematically sorted out, focusing on the method that has the potential for atomic-scale additive manufacturing of 2D materials. Insights into the existing methods to pave the way for advancements in atomic-scale additive manufacturing of 2D materials are provided.

Site-selective CVD method

CVD is widely used for synthesizing large-scale, high-quality, and controllable layers of 2D materials. In the CVD process, the precursor molecules undergo chemical reactions in the vapor phase when exposed to elevated temperatures within a reaction chamber. Following the successful synthesis of graphene^26,27,28, CVD has facilitated the sequential accomplishment of synthesizing various other 2D materials, such as TMDs^29,30,31, h-BN^32,33,34, etc.

Researchers have focused on controlling the growth parameters, such as precursors, temperature, chamber pressure, gas flow rate, and so on, to control the number of layers and the size, morphology, and orientation of the deposited film³⁵. In CVD, precursors are gases or vapors, through chemical reactions within a chamber, synthesize a desired material. Researchers carefully select precursors based on factors like properties, reactivity, and compatibility with CVD conditions. These precursors supply the necessary elements for the target material, such as carbon for graphene or molybdenum and sulfur for MoS₂. For example, in the CVD process for graphene, methane (CH₄) is commonly chosen as the precursor³⁶; for MoS₂, precursors may include molybdenum trioxide (MoO₃) or molybdenum chloride (MoCl₅) as the molybdenum source, along with hydrogen sulfide (H₂S) or sulfur power as the sulfur source³⁷. The selection of a substrate has a direct impact on the initiation of nucleation and subsequent growth, influencing the quality and properties of the resulting material. Certain substrates, such as copper for graphene, possess catalytic properties, contributing to the effectiveness of the overall process³⁵. In CVD processes, elevated temperatures are commonly employed (e.g., approximately 1000 degrees for graphene³⁸ and over 700 degrees for MoS₂³⁹). This is a crucial parameter for the success of chemical reactions that encompass precursor decomposition, nucleation, and crystal growth, and insufficient temperature can lead to low crystalline quality. Pressure and gas flow rate are also critical parameters in the CVD process. Lower pressure is preferred in the CVD process. Referring to the ideal gas equation PV = nRT (P is pressure, V is volume, n is the amount of substance, R is the ideal gas constant, and T is temperature), when operating at low pressure, both the volume flow and gas velocity experience substantial increases for the same molar flow. Concurrently, the precursor concentration decreases. The combination of low concentration and high velocity in the mass feed of the precursor contributes to enhanced controllability of the reaction³⁵. After decades of effort, the CVD method has become a widely used technique for growing 2D materials and can obtain single-layer, large-area wafer-level 2D material films. Gao et al. employed CH₄ as the precursor, along with Ar and H₂ as carrier gases. They successfully produced large-size (4-inch) graphene single-crystal wafers using the CVD process on a single-crystal Cu(111) substrate at a temperature of 1000 °C²².

Some variants of CVD, such as Metal-Organic Chemical Vapor Deposition (MOCVD) and Plasma-Enhanced Chemical Vapor Deposition (PECVD), can also play a significant role in few-layer 2D material. MOCVD is a kind of CVD method that employs organo-metallic precursors. In MOCVD, volatile metal-organic substances in the gas phase are conveyed to the substrate via a carrier gas, where they are integrated into the developing film. MOCVD can more accurately control the supply of precursors and is very promising in steadily growing two-dimensional films with high uniformity on large wafers. Zhu et al. devised a MoS₂ MOCVD approach, employing molybdenum hexacarbonyl (Mo(CO)₆) and diethyl sulfide ((C₂H₅)₂S) as precursors. NaCl acts as a promoter, and Argon functions as the carrier gas. The reaction chamber is divided into two distinct sections: the decomposition area (Area II) with a central temperature of 800 °C, ensuring complete sulfur precursor decomposition, and the growth area (Area I) is maintained below 300 °C. This precise control over temperature and precursor conditions results in the successful fabrication of a single layer of MoS₂ on a 200 mm wafer²³. PECVD utilizes plasma, a state of matter composed of charged particles like electrons and ions, introduced into the reaction chamber alongside precursors to enhance the deposition process. The plasma efficiently decomposes the precursor, aiding nucleation and thin film growth at lower temperatures. Zhang et al. utilized PECVD to deposit a few layers of PdS₂, a kind of TMDs⁴⁰. Pd was sputtered onto a SiO₂/Si substrate, and sublimed sulfur powder served as the precursor. Highly pure N₂ played a dual role as both the plasma feed gas and carrier gas. The plasma discharge, induced by an RF generator, operated at a maximum temperature of 475 °C. This PECVD process allowed for efficient and controlled deposition of PdS₂ layers.

These studies have predominantly succeeded in depositing a single or a few layers of 2D materials and have taken a solid step for 2D materials to move out of the laboratory and into the industry. For the next step, to achieve CVD-based additive manufacturing of 2D materials, a crucial requirement is the precise selection and synthesis of materials at the desired positions. Site-selective CVD method mainly includes two types: seed method and substrate pre-patterning method. The two methods will be introduced in the following sections separately.

Seed method

In the typical CVD procedure for depositing 2D materials, grain nucleation occurs at random locations, and these grains gradually merge to form a complete film. The seed method initially aimed to achieve small-sized, high-quality single-crystal 2D materials without grain boundaries rather than site-selective growth. However, it has proven to be an effective approach for achieving both objectives. The use of seeds allows for precise control over the position of crystal nucleation. Figure 2d illustrates the CVD synthesis of graphene grains with and without the use of seeds. In the area without seeds, random positions of nuclei are observed, resulting in a non-uniform distribution. On the other hand, in the region where seeds were strategically placed in advance, graphene grains form an orderly matrix. This phenomenon can be explained by the principle that seeds provide a lower energy barrier, facilitating controlled nucleation at specific sites⁴¹. Seeds can be raw materials^42,43,44, metal^45,46 and defects on substrates⁴⁷, which are summarized in Table 1.

Fig. 2: Seed methods for site-selective CVD of 2Dmaterials.

a SEM image showing an array of graphene seed patterned on Cu foil. b SEM image of graphene grain array grown from an array of seed crystals after 5 min growth, many grains grow from the seeds, while there is a grain that nucleated randomly at the lower left. c SEM image of graphene grain array grown from an array of seed crystals after 15 min growth. d SEM image of graphene growth with and without seed method (two areas are divided by dotted line in the figure) Reproduced from ref. 42 with permission from Springer Nature, Copyright 2011. e MoS2 monolayers grown from the patterns of Au seeds array. f Optical image of the MoS2 monolayers line used as the device channel by designing the Au seed arrangement. g Optical image of multiple MoS2 transistors by pattern seeds. Reproduced from ref. 46 with permission from American Chemical Society, Copyright 2018. h MoS2 CVD process seeded with holes: cover a 300 nm SiO2/Si wafer with a chromium film and photoresist; expose the resist in a mask aligner and develop it; use wet etching to transfer the hole pattern onto the chromium layer; remove the remaining resist and use the chromium layer as a hard mask for inductively coupled plasma etching of the oxide; wet remove the chromium layer and clean the substrate with piranha solution; seed MoS2 growth at the holes. Reproduced from ref. 47 with permission from IOP Publishing, Copyright 2014.

Table 1 Representative studies on seed method for site-selective CVD of 2D material

In 2011, Yu et al. reported the seeded growth method. The researchers successfully demonstrated the synthesis of single-crystal graphene grains with about 10 micrometers in size array on polycrystalline Cu using predetermined graphene seeds. The method involves a two-step process utilizing CVD⁴². In the first step of the CVD process, multilayer graphene is grown on a Cu foil. Subsequently, the lithography method is employed to create an array of seed crystals of graphene on the Cu foil, shown in Fig. 2a. In the second step of the CVD, the graphene grains are grown specifically at the predetermined seed locations. This results in the formation of a well-defined graphene grain array, as illustrated in Fig. 2b and c. Wu et al. introduced the use of electron beam lithography to pattern polymethyl methacrylate (PMMA) onto annealed Cu foils, creating seed arrays⁴³. The inclusion of PMMA seeds serves to enhance the carbon concentration in specific regions, thereby facilitating graphene nucleation at the corresponding sites during the CVD process. It eliminates the need for an extra CVD process by directly introducing the seed arrays onto the Cu foils using electron beam lithography.

The seed method has been extended to other 2D materials, such as MoS₂. Han et al. successfully grew flakes of monolayer MoS₂ at predetermined locations on SiO₂ by utilizing a micrometer-scale resolution patterned MoO₃ seed array⁴⁴. The MoO₃ seeds were fabricated using photolithography followed by either thermal evaporation and following lift off or an aqueous solution of ammonium heptamolybdate (AHM) with a specialized process. After pretreatments, the sample with MoO₃ seeds was placed in a furnace for vulcanization. An array of triangular/spike-shaped flakes of monolayer MoS₂ was successfully grown on SiO₂. In the method, growing MoS₂ on a SiO₂/Si substrate eliminates the need for transfer steps, which was required in graphene production on Cu foil, thus demonstrating its compatibility with conventional manufacturing processes.

Metal as a seed for 2D material site-selective growth has also been reported. Miseikis et al. patterned a chromium array on natively-oxidised copper foils via optical lithography and thermal evaporation as seeds, enabling the growth of graphene patterns⁴⁵. Li et al. employed electron-beam lithography to prepattern arrays of gold (Au) nanostructures on a SiO₂ substrate to achieve selective-area growth of MoS₂ monolayers, as shown in Fig. 2e⁴⁶. The mechanism involved the deposition of MoS₂ layers on the surface of Au nanoparticles or the interface of Au–SiO₂ initially, followed by a classic heterogeneous nucleation process and lateral layer growth⁴⁶. Through the design of the Au seed arrangement, a MoS2 line is grown as the channel of a typical field-effect transistor (FET), shown in Fig. 2f, g. It should be noted that in the metal seeding method, the metal seeds are preserved in the product, which may have an impact on the performance of the resulting devices. In addition, defects on the substrate can also act as seeds. Sun et al. patternedΦ= 2 µm round holes by the typical photolithography etching method into a SiO₂/Si substrate at a pitch of 7 µm as seed, then grow MoS₂ by CVD⁴⁷. The holes also affect the growth position of the MoS₂ film, realizing site-selective growth. The specific process is shown in Fig. 2h.

The seeding method offers advantages by leveraging well-established CVD techniques and facilitating the production of high-quality films of 2D materials. However, it has limitations regarding precise control over the growth direction and range of the 2D materials. The fabrication of the seed structures plays a crucial role in determining the line width of the resulting 2D material patterns. Imperfections or variations in the seed fabrication process can directly affect the width and quality of the grown patterns. Another concern with the seeding method is the retention of metal and defective seeds within the 2D material structure. These seeds can impact subsequent processing steps and device performance.

Pre-patterning of the substrate

Pre-patterning the substrate serves as an alternative approach for achieving area-selective growth of 2D materials via CVD. This method involves patterning the substrate using various materials, including metals or metal oxides. By leveraging the inherent selectivity of the 2D material towards the growth interface or by directly converting the patterned substrate material into the desired 2D material, site-selective growth can be achieved. Furthermore, researchers have proposed non-destructive pattern processing techniques for the substrate to enable site-selective growth. A summary of representative studies in this field can be found in Table 2.

Table 2 Representative studies on prepatterning substrate for site-selective CVD of 2D material

Kim et al. used the natural properties of graphene and synthesized patterned graphene on the pre-patterned nickel layer on SiO₂ substrate. The underlying principle is based on the fact that the reactive nickel surfaces facilitate the growth of thin graphitic layers through CVD, while SiO₂ lacks this capability⁴⁸. Xue et al. used lithography and electron beam evaporation to pre-deposit an array of WO₃ and Mo precursors on a SiO₂/Si substrate, which were then sulfurized into MoS₂ and WS₂ array via an ambient pressure thermal reduction process⁴⁹. Woods utilized a similar approach, involving sequential magnetron sputtering and electron beam lithography to pattern Mo and W on SiO₂/Si substrate as precursors, which were subsequently sulfurized to form a MoS₂/WS₂ horizontal heterojunction, shown in Fig. 3a⁵⁰. This method offers a convenient means for creating heterojunctions; however, it requires prior deposition and patterning of a metal or its oxide.

Fig. 3: Pre-patterning substrate for site-selective CVDof 2D materials.

a Schematic of MoS₂/WS₂ growth from patterned Mo and W layers. Reproduced from ref. 50 with permission from American Chemical Society, Copyright 2016. b Schematic diagram of the process of O₂ plasma treatment of SiO₂ and CVD of MoS₂ on bare (process I) or graphene-covered (process II) SiO₂/Si wafers. In process II graphene exposed of O₂ plasma was removed, and MoS₂ grows between graphene. Reproduced from ref. 51 with permission from American Chemical Society, Copyright 2009. c, d schematic illustration of a rubbing tool to generate triboelectric charge patterns on substrate to induct MoS₂ selective growth. e illustration of MoS₂ CVD growth based on rubbing-Induced method. Reproduced from ref. 53 with permission from American Chemical Society, Copyright 2018.

Plasma treatment is known to induce various surface modifications, and it can also play a role in selectively synthesizing 2D materials. Chen et al. developed a lithography-free plasma-induced patterned method for the selective growth of MoS₂⁵¹. The method involves treating the SiO₂/Si substrate with O₂ plasma 30 second under a shadow mask, which leads to the selective growth of MoS₂ in the treated regions during CVD deposition, shown as the process I in Fig. 3b. Additionally, when graphene was first transferred onto the SiO₂ substrate and treated with O₂ plasma, only the masked graphene remained during the treatment. MoS₂ could then grow on the treated surface between two neighboring graphene layers, resulting in the direct formation of a MoS₂/graphene heterojunction, shown as process II in Fig. 3b. Guo et al. also utilized a combination of photolithographic methods and O₂ plasma to selectively grow monolayer MoS₂ on SiO₂⁵². Initially, the pattern was defined through photolithography, followed by O₂ plasma treatment of the exposed surface. After the resist removed, CVD process was applied, resulting in selective MoS₂ growth only at the O₂ plasma-treated regions with hydrophilic salt perylene-3,4,9,10-tetracarboxylic acid tetrapotassium (PTAS) used for promoting seeded. The selectivity achieved through O₂ plasma treatment stems from its capability to elevate surface energy and improve hydrophilicity. For instance, on a SiO₂/Si substrate, the initial contact angle is approximately 38.48°. However, after a 10-minute O₂ plasma treatment, the contact angle significantly decreases to 6.04°. Leveraging hydrophilic salts in conjunction with this treatment enabled the synthesis of MoS₂ with a width resolution of 2 μm, spanning from a single layer to a few layers. Other lithography-free method is reported, Ryu et al. proposes a rubbing-Induced method⁵³, which involves rubbing a period Cu-coated Si rubbing template against SiO₂ surfaces under a contact force of 100 mN to create a patterned surface potential on a damage-free SiO₂ substrate, shown in Fig. 3c, d. MoS₂ is then selectively deposited onto the rubbing-induced areas using CVD deposition (Fig. 3e). Such lithography-free substrate patterning methods are attractive, but the resolution of these methods is still relatively low.

In comparison to the seed method, the substrate pretreatment method offers improved control over the growth range of 2D materials, and it results in 2D material films without residual seeds. However, the success of the substrate prepatterning method relies heavily on the photolithography technique. Although recent advancements have introduced nondestructive methods to avoid the use of photolithography, there is still a need for further improvements in achieving higher resolution in the selected areas.

Area-selective ALD

Atomic Layer Deposition (ALD) is a thin-film deposition technique that develops from CVD. In binary reaction CVD, such as A reacts with B to generate a product, both A and B reactants coexist concurrently throughout the deposition process, resulting in the continuous formation of the product film on the substrate. In contrast, ALD employs a stepwise approach where the substrate is exposed separately to reactants A and B. To design an ALD process, a CVD process based on a binary reaction should be found. Subsequently, the A and B reactants are introduced sequentially in the binary reaction sequence, enabling the precise and controlled layering required for the target deposition^54,55. Due to the characteristics of ALD, it emerges as a promising technique for the atomic-scale fabrication of two-dimensional materials. One notable advantage of ALD is its ability to operate at relatively low temperatures (<350 °C), in contrast to the high-temperature environment (over 800 °C) typically required by conventional CVD processes⁵⁶. This lower temperature requirement in ALD not only enhances its compatibility with a wider range of substrates but also minimizes thermal stress and enables the synthesis of high-quality 2D materials. ALD has demonstrated its capability to synthesize 2D materials, such as graphene, h-BN, and TMDs^25,57. Area-selected ALD and site-selective CVD both use seed methods⁵⁸ and patterned substrates^59,60,61. Moreover, the use of inhibitors on patterned surfaces has been demonstrated to be an effective means of achieving selective growth for ALD^62,63. These advancements and their corresponding methodologies are summarized in Table 3.

Table 3 Representative studies on area-selected ALD of 2D materials

Groven et al. placed WS₂ seeds at specific locations on Si substrate, then PEALD (plasma-enhanced atomic layer deposition) is applied to obtain a WS₂ film⁵⁸. The precursors employed in this process include WF₆ and H₂S. H₂ plasma serves to reduce the adsorbed WF_x species initially in the VI+ oxidation state. This reduction enables subsequent WF₆ adsorption by creating reactive surface sites around the WS₂ seed, achieving a well-controlled composition and lateral growth. The lateral crystal size of the resulting film was manipulated by adjusting the number of PEALD reaction cycles, while the film’s thickness was contingent on the thickness of the seeds. Jurca et al. proposed a process combining lithography with area-selective ALD to produce MoS₂ films in specific locations on a Si/Si₃N₄ substrate; the process is shown in Fig. 4a⁵⁹. The low-temperature ALD process used H₂S and Mo(NMe₂)₄ as precursors and could be performed at 60°C. First, the photoresist was coated and patterned using lithography to define the desired locations. Then, ALD was applied. After removing the photoresist, the selectively grown MoS₂ film is obtained.

Fig. 4: Area-selected ALD of 2D materials.

a ALD process based on a resist grows patterned MoS by either maskless photolithography or electron-beam lithography. Reproduced from ref. 59 with permission from John Wiley and Sons, Copyright 2017. b Area-selective ALD process of MoS₂ on different surfaces (SiO₂ and Al) based on self-etching by prolonged MoCl₅ injection. Reproduced from ref. 62 with permission from John Wiley and Sons, Copyright 2021. c Area-selective ALD process for WS₂ using ABC-type ALD cycles based on inhibitor molecules. Reproduced from ref. 63 with permission from American Chemical Society, Copyright 2020.

Yue et al. reported a method for selectively growing MoS₂ on patterned Au pads on a SiO₂ substrate using molybdenum chloride (MoCl₅) and hexamethyldisilathiane (HMDST) as precursors⁶⁰. The selectivity in this process arises from the high catalytic reactivity of Au with these ALD precursors. Sharma et al. published an approach involves patterned PEALD deposition of MoO_x as the parent material using electron beam lithography (EBL)⁶¹. The PEALD process is carried out at a low temperature of 50 °C. The resulting MoO_x thin film is then sulfurized at a high temperature of 900 °C using a mixture of H₂S and Ar gas to obtain a MoS₂ film. This approach allows for direct growth of the film on device-ready substrates without the need for transfer. Additionally, WS₂ can also be synthesized using a process similar to that of PEALD-deposited WO_x, and MoS₂-WS₂ heterostructures have also been demonstrated using this method.

Ahn et al. proposed an approach to grow MoS₂ on predefined Al patterns on a SiO₂ substrate, using MoCl₅ and H₂S as precursors, shown in Fig. 4b⁶². MoS₂ can deposit on both Al and SiO₂ surfaces under appropriate precursor concentrations. However, the highly reactive MoCl₅ can induce a self-etching side-effect, and the side-effect is more sensitive on SiO₂ surfaces. Therefore, deposition of MoS₂ on SiO₂ surface can be inhibited under higher MoCl₅ concentration, whereas the inhibition effect on the Al surface is not as significant. As a result, the MoS₂ is deposited selectively on predefined Al patterns rather than on the SiO₂ surface, and no selectivity loss was observed even up to 400 cycles ALD. Shashank et al. employed pre-patterned Al₂O₃ (non-growth area) and SiO₂ (growth area) surface to grow WS₂ film in an ABC-step PEALD process, shown in Fig. 4c⁶³. In step A, inhibitor molecules (acetylacetone) were selectively adsorbed on the non-growth area but not on the growth area. In step B, the tungsten precursor was absorbed by the SiO₂ surface while its adsorption on the non-growth site was blocked by the adsorbed inhibitor molecules. In step C, H₂S plasma was used as the ALD co-reactant to enable the growth of WS₂ on SiO₂, and the process was repeated. They also compared WS₂ growth on different transition metal oxide surfaces using the same ABC steps. The results showed a growth delay at the beginning loop for WS₂ growth on Al₂O₃ and HfO₂, and the growth speed was very slow. Immediate film growth was observed on MoO₃ and Nb₂O₅ surfaces.

The exploration of area-selective ALD continues to advance. While some interesting results have been achieved, it faces a similar challenge as site-selective CVD in its reliance on traditional photolithography processes. To enable true additive manufacturing, it is crucial to minimize the use of such processes. Currently, there is limited research on non-lithographic methods for area-selective ALD of 2D materials. However, Kundrata et al. proposed a potential solution using microfluidic gas-delivered ALD for atomic-scale additive manufacturing⁶⁴. Their approach involves a miniaturized concentric nozzle to precisely deliver the gas precursor to the desired location, along with vacuum channels to prevent gas diffusion and ensure high resolution. While this technique has not been specifically tested with 2D materials, it presents a flexible approach to achieve selective region fabrication in ALD. It holds promise for future advancements in non-lithographic area-selective ALD techniques.

Electrodeposition

Electrodeposition, a technique rooted in electrochemistry, has found widespread applications in the semiconductor industry. The process entails the controlled reduction of metal ions from a solution onto a substrate, facilitated by an electric field, offering precise control over film thickness and morphology. Electrodeposition configurations for 2D material generally include two electrodes or three electrodes. The two-electrode configuration is simple, including a working electrode (substrate) and a counter electrode, but it faces challenges in controlling the working electrode during operation. Consequently, the three-electrode configuration addresses this issue by introducing a reference electrode, enabling the measurement and control of the working potential. Electrodeposition governs material deposition by manipulating parameters like electrolyte composition (precursor), operating voltage, deposition time, and others. Before electrodeposition, cyclic voltammetry (CV) scanning can be conducted to investigate the electrochemical behavior of the electrode and electrolyte, aiding in the determination of the optimal operating voltage. Electrodeposition methods have been used for synthesizing two-dimensional materials, including graphene^65,66,67, h-BN⁶⁸, TMDs^69,70,71 and 2D material-based nanocomposites^72,73. One notable advantage of electrodeposition is its ability to operate at room temperature and atmospheric pressure, eliminating the need for costly equipment. By utilizing pre-patterned electrodes, area-selective growth of 2D materials can be achieved through electrodeposition. Current advancements in this field have primarily focused on MoS₂ and WS₂, and the precursors used are mainly [MoS₄]²⁻ for MoS₂^74,75,76,77 and [WS₂Cl₂]^2- for WS₂^78,79, as summarized in Table 4.

Table 4 Representative studies on area-selective electrodeposition of 2D materials

Wan et al. successfully produced a MoS₂/graphene heterostructure through electrodeposition using (NH₄)₂MoS₄ as precursor and water as electrolyte⁷⁴, the structure of the experiment is shown in Fig. 5a. A monolayer graphene film on a SiO₂/Si substrate is used as the working electrode, with a carbon rod serving as the counter electrode. A constant-current source was used due to the variations in the circuit’s total resistance caused by the concentration of the electrolyte and other factors. The first step involved the oxidative electrodeposition of a MoS₃ thin film on the graphene film, with the reduction at the counter electrode producing MoS₂, shown in Fig. 5b, c. Both electrodes were then annealed to obtain a MoS₂ film at a temperature of around 500–800 °C. The thickness of the deposited MoS₂ film on the graphene film working electrode was estimated to be between 5 and 150 nm, which can be controlled by changing the concentration of the precursor, current density, and deposited time. Electrodeposition of 2D materials based on an aqueous solution is feasible. However, using water as a solvent for electrodeposition may lead to some problems. Water can also be electrolyzed; to avoid it, the electrochemical window is limited. Possible hydrogen evolution on the electrode may affect the quality of the deposited film⁷⁵. Electrodeposition methods based on non-aqueous solvents are necessary.

Fig. 5: Area-selective electrodeposition of 2D materials.

a Schematic of the two electrodes set system for electrodeposition MoS₂ on graphene. b After electrodeposition, an annealing step is applied. c The reaction on the two electrodes. Reproduced from ref. 74 with permission from John Wiley and Sons, Copyright 2017. d (left) Schematic of the three electrodes set system for electrodeposition MoS₂ on graphene; (right up) the final devices and (right bottom) annular dark-field (ADF) TEM images of electrodeposited MoS₂ film. Reproduced from ref. 76 with permission from American Chemical Society, Copyright 2020. e The illustration of MoS₂ film growing laterally on the SiO₂/Si substrate starts from TiN electrodes covered with a SiO₂ insulator. f the SEM image of four TiN electrodes and three micro-gaps filled with laterally grown MoS₂ films. Reproduced from ref. 77 with permission from American Chemical Society, Copyright 2022.

Noori et al. reported using tetrabutylammonium tetrathiomolybdate [NnBu₄]₂[MoS₄] as a single-source precursor and nonaqueous solvent dichloromethane (CH₂Cl₂) for electrodepositing MoS₂ on graphene⁷⁶. Three electrode configurations were used, including graphene as the working electrode, Pt gauze as the counter electrode, and Ag/AgCl as the reference electrode, shown in Fig. 5d. A sealed container was used to ensure the deposition of MoS₂ in a selective area. [MoS₄]^2- was reduced to MoS₂ on the working electrode, and the thickness of the deposited MoS₂ could be controlled by the deposition time. After 90 seconds of deposition, the thickness was measured to be 5.8 nm. An annealing process is necessary for electrodeposition since the deposited MoS₂ was amorphous. Abdelazim et al. utilized the same precursor and solvent to achieve lateral growth of MoS₂ on a SiO₂/Si substrate⁷⁷. They fabricated 100 nm thick TiN electrodes using a photolithographic process and sput method onto the SiO₂/Si substrate as the working electrode, shown in Fig. 5e. [MoS₄]²⁻ was deposited on the working electrode and reduced to MoS₂. The lateral growth started from the side edges of the TiN electrodes and grew on the SiO₂/Si substrate, as shown in Fig. 5f. The rate of lateral growth was found to be linear at approximately 33 ± 6 nm per min, and was approximately 20 times faster than the rate of vertical growth. This shows that this method has good control over the growth direction. The same research group also conducted an electrodeposition of WS₂ on a patterned graphene substrate, resulting in the selective growth of patterned layers of WS₂⁷⁸. The precursor used was [NEt₄]₂[WS₂Cl₄], while CH₂Cl₂ was employed as a solvent. The patterned graphene was fabricated through a process of UV lithography and reactive ion etching (RIE) as the working electrode. The findings indicated that WS₂ exclusively grew on the patterned graphene, enabling the selective growth of WS₂. The thickness of the deposited WS₂ was controlled by varying the deposition time, resulting in thicknesses of 25 nm, 7.2 nm, and 2.1 nm for deposition durations of 5 min, 1 min, and 10 s, respectively.

Nanocomposite films can also be synthesized using the electrodeposition technique. The incorporation of 2D materials into a matrix or host material results in the formation of composite structures known as 2D material-based nanocomposite films. This process is similar to the typical electrodeposition, except for the inclusion of 2D materials into the electroplating solutions. These films amalgamate the properties of 2D materials with the bulk properties of the matrix material, leading to enhanced characteristics. For instance, a nano-Ni/reduced graphene oxide composite film synthesized through electrodeposition exhibits robust electroactivity, making it well-suited for high-precision sensors based on electrochemistry^72,73. Co₃O₄-graphene nanocomposites prove highly suitable for fabricating supercapacitors⁸⁰. Similar to conventional electrodeposition approaches, the electrodeposition of nanocomposite films with 2D materials offers the potential to achieve nearly atomically thin films. This can be achieved through meticulous control of precursors, electrode setup, controlled growth time, and other influencing factors. It is important to note that the majority of existing nanocomposite films based on 2D materials and electrodeposition exhibit complex 3D morphologies. As of now, they have not yet reached the level of atomic layer thinness.

Electrodeposition of 2D materials offers several advantages; it allows for the formation of patterned 2D materials and vertical heterojunctions by utilizing pre-patterned electrode materials. Studies on electrodeposition have demonstrated a linear deposition rate⁷⁸, enabling precise control over the thickness of the deposited layer at the atomic level by adjusting the deposition time. The synthesis of 2D materials with complex 3D structures through electrodeposition methods is possible. The template-assisted 3D electrodeposition method, exemplified by its success in creating 3D spring structures using nanocrystalline (NC) nickel⁸¹, emerges as a promising solution. In this approach, a template serves as a mold or scaffold, guiding the deposition of material onto a substrate to shape intricate 3D structures. Through this method, 2D materials can be controllably and efficiently synthesized into complex structures that may be challenging to achieve directly with alternative additive manufacturing methods. While the method faces limitations in achieving atomic-scale accuracy due to template size constraints, it holds substantial potential within the N/MEMS field. However, there are certain limitations associated with electrodeposition. The quality of the deposited 2D material can be affected by impurities in the electrolyte. Moreover, successful electrodeposition relies on suitable redox reactions of precursors, which means that not all 2D materials can be synthesized using this method. Additionally, the 2D material films generated by electrodeposition are typically amorphous, necessitating an additional annealing step to enhance their crystallinity.

Laser-assisted method

Laser direct writing is a versatile technique that harnesses the focused energy of lasers to create micro-nano structures on various materials without direct contact; this method has gained widespread adoption. In Additive Manufacturing, it also played an important role^82,83. One of the key advantages of laser direct writing is its capability of producing highly precise structures without the need for traditional photolithographic etching processes. This eliminates the risk of surface contamination and material damage. 2D materials can be synthesized through laser-assisted techniques, wherein a typical procedure involves applying a solution with precursors onto the desired substrate via spin-coating. Subsequently, laser irradiation is employed to decompose and transform the precursors, achieving the patterning of several layers through thermal effects^84,85,86,87 and photochemical reactions^88,89,90, where the detailed information is summarized in Table 5. Notably, this synthesis of 2D materials can be accomplished at room temperature.

Table 5 Representative studies on the laser-assisted method of 2D materials

Laser-assisted techniques have been applied in the fabrication of two-dimensional (2D) materials, starting with graphene. Lin et al. used CO₂ infrared laser with 10.6 µm wavelength and 14 μs pulses to irradiate an insulating polyimide (PI) film, resulting in the generation of porous graphene⁸⁶. The localized high temperatures (up to 2500 °C) produced by the laser broke the existing chemical bonds and formed the graphene structure. However, the method did not allow for control over the number of graphene layers.

Laser-assisted manufacturing of 2D materials with controlled layers has mainly focused on TMDs, such as MoS₂ and WS₂. Hu et al. used a 248 nm wavelength krypton fluoride laser to deposit a MoS₂ line on a Si/SiO₂ substrate in a vacuum environment⁸⁹, the schematic diagram is shown in Fig. 6a. The precursor used was (NH₄)₂MoS₄, and it was irradiated by laser with the energy density 250 mJmm^-2 for 5–10 min to thermal decompose to MoS₂⁹¹. The specific operations are shown in Fig. 6b. By changing the pulse number, the MoS₂ structure can be controlled to realize the structure of 1 T, 2H and coexistence, shown in Fig. 6c. The thickness of the MoS₂ thin film can be controlled by the concentration of the precursor, and at lower concentrations, the generated thinnest MoS₂ reached a thickness of approximately 6.09 nm.

Fig. 6: Laser-assisted method of 2D materials.

a Schematic of the laser-induced synthesis experimental device. b Typical steps of laser-induced synthesis experiment, including coating, laser irradiation and cleaning. c Under the action of laser (MoS4)^2- synthesise MoS₂ structure with coexistence of 1 T and 2H. Reproduced from ref. 89 with permission from Royal Society of Chemistry, Copyright 2012. d Through two cycles, MoS₂ and WS₂ are grown by laser-assisted thermal decomposition on different positions on the same substrate. Reproduced from ref. 84 with permission from American Chemical Society, Copyright 2020. e, g Schematic illustration of the synthesis of MoS₂ by femtosecond laser direct writing. f photochemical reaction induced by TPA. Reproduced from ref. 88 with permission from American Chemical Society, Copyright 2022.

Jung et al. improve the laser-assisted method of MoS₂ synthesis at room temperature and pressure with the precursor (NH₄)₂MoS₄⁹⁰. The process includes coating precursor evenly on substrate, solidifying baking and irradiation with a laser of 532 nm wavelength. The study found that the minimum power required for synthesis is 3 mW, and the synthesis of the MoS₂ film takes only a few microseconds under exposed laser. Additionally, this method can also deposit MoS₂ on transparent flexible PET substrates, providing a solution for flexible sensing. Different from the previously mentioned thickness of MoS₂ formed is controlled by the precursor concentration, in this study, they found the thickness of the film can be controlled by adjusting the laser power, with higher laser power resulting in thinner films. The thinnest film obtained was 4 nm. This rationale is justified by the dual-step nature of the process. In the initial step, the laser facilitates the photochemical reaction of the precursor, resulting in the formation of MoS₂. Subsequently, in the second step, the heating induced by light absorption causes the sublimation of the upper layers⁹². Notably, as the laser power increases, a greater quantity of synthesized the MoS₂ is removed, culminating in the attainment of a thinner film through higher laser power.

Park et al. have developed a technique for depositing layers of MoS₂ and WS₂ using a pulsed laser with a wavelength of 1.06 µm and a pulse duration of 100 ps⁸⁴. The method is based on thermal decomposition, with the key factor being the heat absorption of the materials. Different materials have varying optical absorption factors, when the laser target a specific point, because of the different optical absorption factors, the laser can heat a specific layer of materials with precision. The precursors for this method are (NH₄)₂MoS₄ and (NH₄)₂WS₄, and if the substrate has a lower optical absorption factor than the precursors, MoS₂ and WS₂ can be deposited on it. The potential substrate can be flexible materials, SiO₂/Si substrate, or other two-dimensional materials, where the steps are shown in Fig. 6d. The pattern width is determined by the laser scribing speed and beam size (laser scribing speed at 5 mm per second, the pattern width was ∼40 µm and at 20 mm per second, the width became nearly comparable to the beam size 20 µm), while the thickness is controlled by the thickness of the precursor layer, the thinnest layer is 2.2 ± 0.4 nm under thicknesses of the precursor films 6.1 ± 0.6 nm.

Xu et al. have proposed an alternative laser method for growing MoS₂ using thermally-induced chemical reactions with a femtosecond laser with 780 nm wavelength⁸⁸, as shown in Fig. 6e. In this method, molybdenum acetylacetonate Mo (ACAC)₂ and carbon disulfide CS₂ are used as precursors. When the laser irradiates the precursors, two-photon absorption (TPA) occurs, inducing a photochemical reaction, which results in the growth of MoS₂. The writing speed of this method can reach up to 150 µm per second, which is significantly faster than the thermal decomposition method. However, the precision in controlling the thickness of the deposited MoS₂ using this method is not as advanced as that achieved through the thermal decomposition method mentioned earlier, the thinnest layer is about 40 nm, the narrowest wide is about 780 nm. Averchenko et al. developed a method to synthesise Mo_1-xW_xS₂ alloys using (NH₄)₂MoS₄ and (NH₄)₂WS₄ solutions as precursors with a 532 nm laser system⁸⁷. By manipulating the laser power and scanning speed during laser direct-write, researchers were able to achieve control over the width and thickness of the alloy lines formed. The narrowest width achieved in their experiments was 3.73 µm. The thickness of the formed lines had two side-lobes, with the sides being thicker than the middle part, and the middle part thickness being 4 nm, which shows a non-uniform distribution of material deposition, accentuated by a stronger laser-thinning effect in the central region during the laser direct-write process.

The Laser Powder Bed Fusion (L-PBF) method, widely employed for printing metal 3D structures, also finds applications in the realm of 2D materials⁹³. In a study conducted by Navid Alinejadian et al., Selective Laser Melting (SLM) technology was successfully utilized to fabricate a MoS₂/Mo₂S₃ nanocomposite^94,95. The process involved the meticulous mixing of gas-atomized powder of pure molybdenum and pure 2H-MoS₂ powder as raw materials. Subsequently, laser scanning was employed to systematically convert the sample into Mo/Mo_(x)S_(x+1) layers, resulting in a MoS₂/Mo₂S₃ nanocomposite. It is noteworthy that the utilization of a high-energy laser in this process generates numerous α-phase nucleation sites in MoS₂, leading to a phase transformation from the semiconducting 2H phase to the metallic 1 T phase. While this manufacturing method demonstrates significant flexibility in creating nanocomposites based on 2D materials, it is essential to acknowledge that the thickness of the layer formed during each laser scan is challenging to control precisely. Consequently, achieving atomic-scale manufacturing remains a formidable task in this context.

The laser-assisted method offers flexibility and speed in obtaining thin layers of 2D materials at room temperature and atmospheric pressure. The thermal effect of the laser enables the formation of well-crystallized 2D material films without the need for additional annealing steps. Additionally, lasers can directly synthesize vertical heterojunctions of 2D materials and can be applied to flexible substrates, making it a promising technique for flexible electronics applications. However, the laser methods present challenges in directly and precisely controlling the thickness of 2D material films, as well as in manufacturing large-scale films. The synthesis of 2D materials using lasers is limited to precursors that can be decomposed through pyrolysis or light-induced decomposition, restricting the range of materials that can be produced using this method.

Printing method

Selective deposition of 2D materials has also been achieved through printing methods, which can be broadly classified into two techniques: stamp printing^96,97, inkjet printing and its variants^{98,99,100,101,102,103,104,105}. Stamp printing involves the transfer of pre-synthesized 2D materials using stamps forming a patterned layer. Inkjet printing involves blending precursors with auxiliary solutions to form ink with the appropriate viscosity and stability. This ink is subsequently dispensed onto the substrate using a printer or similar printing technology and is transformed into the desired 2D materials through heat treatment. These methods are summarized in Table 6 and will be discussed in detail below.

Table 6 Representative studies on printing method of 2D materials

Stamp printing offers precise and controlled transfer of 2D materials onto specific substrates. The method has high positioning accuracy, potentially in the nanometer range. The stamp printing method, also known as the transfer printing process, shares similarities with the conventional stamp-based transfer process. The distinction lies in the materials and the design of the working surface of the stamp of both. In the dry transfer method, the stamp is always made of polymer and the stamp’s working surface is a flat plane, typically employed for transferring a single layer or a few layers of irregularly shaped 2D material flakes. Conversely, in stamp printing, the working surface of the stamp is not a uniform plane; rather, it features specific shapes such as circular or rectangular matrices. With stamp printing, patterned 2D material layers are precisely transferred onto the target substrate. The stamp printing method for graphene was demonstrated in 2007 by Liang et al., who used a SiO₂/Si stamp with a pillar array of 15 μm diameter to exfoliate and transfer graphene onto a given region of a substrate⁹⁶. In this method, a layer of thin “transferring” material is attached to the stamp for cutting the graphene, and a layer of “fixing” material is attached to the substrate to provide adhesion between the substrate and graphene and to help the graphene transfer from the stamp. The successful implementation of this method resulted in the precise transfer of a graphene array, mirroring the shape of the stamp, with a thickness of approximately 10 nm, onto the substrate. Nam et al. further developed the stamp printing method to transfer MoS₂ onto a substrate via Plasma-Assisted⁹⁷. In this approach, prepatterned relief structures on the bulk MoS₂ film, created through photolithography and plasma etching, served as the stamp, as shown in Fig. 7a. This contrasts with using a separate material for the stamp. The SiO₂/Si substrate underwent O₂ plasma treatment to introduce uniformly distributed electric charges on the surface, shown in Fig. 7b, enhancing adhesion through additional electrostatic attractive stress. During the stamping process, when the relief structures of the stamp contacted the substrate, prepatterned few-layer MoS₂ flakes are transferred onto the substrate, as shown in Fig. 7c. The resulting MoS₂ thickness was measured to be 3 ± 1.9 nm. Although fewer layers of 2D materials can be obtained through this method, it is obvious that the thickness consistency control is insufficient.

Fig. 7: Printing method of 2D materials.

a–c Schematic of stamp printing method of MoS₂ array (a) prepatterned relief structures on the bulk MoS₂ film; (b) O2 plasma treatment for the SiO₂/Si substrate; (c) Printed few-layer MoS₂ flake array. Reproduced from ref. 97 with permission from American Chemical Society, Copyright 2013. d Schematic of synthesising MoS₂ patterns on flexible substrates, including two main steps: inkjet printing and annealing. Reproduced from ref. 102 with permission from John Wiley and Sons, Copyright 2022. e Schematic of aerosol-jet printing setup: aerosolized ink, carried by a gas stream, is jetted through a nozzle to produce patterns on the substrate. Reproduced from ref. 101 with permission from Elsevier, Copyright 2015. f nozzle and substrate configuration for EHD printing. Reproduced from ref. 106 with permission from Springer Nature, Copyright 2007.

While stamp printing allows for the rapid production of 2D material films with specific shapes, it may not be as flexible for use in additive manufacturing due to its transfer-based nature. In contrast, inkjet printing offers a more versatile solution. McManus et al. developed a water-based ink containing 2D materials, optimised for film deposition⁹⁸. The ink was obtained through liquid phase exfoliation of the 2D material, resulting in a printed product consisting of multilayered structures rather than single-layer atomic films. Wan et al. demonstrated the use of an industrial inkjet printer to print monolayer MoS₂ and MoSe₂ on SiO₂/Si substrate using aqueous precursors⁹⁹. To achieve ink formulation, they modified the surface tension and viscosity of the water solution of ammonium molybdate tetrahydrate ((NH₄)₆Mo₇O₂₄·4H₂O) to obtain ink with desired properties. The ink is printed by a customized inkjet printer. After printing, the substrate and the printed patterns were rapidly moved to a hot zone, where Ar gas and high temperature were applied to generate MoO₃, the MoO₃ was then reacted with S or Se to form MoS₂ or MoSe₂. This inkjet printing method provides a more flexible solution for depositing 2D materials and allows for precise control over the size and location of the printed patterns. Li et al. use of printing-based 2D materials in the field of flexible electronics¹⁰², they choose ammonium thiomolybdate ((NH₄)₂MoS₄) as the precursor and optimize the precursor ink. By inkjet printing, the ink is printed on the flexible substrate polyimide films according to the pre-set pattern, then baking it at 110 °C for 10 minutes to dry the ink, following annealing step at 350–500 °C MoS₂ layer is formed, as shown in Fig. 7d. At a low concentration of the ink, the ultrathin layer can be achieved, the thinnest layer is about 1.43 nm.

To enhance the width resolution of printed lines, various printing techniques are under investigation. Electrohydrodynamic (EHD) jet printing stands as a compelling avenue for achieving high-precision printing, employing electric fields to manipulate liquid droplets instead of relying on mechanical pressure, as shown in Fig. 7f. In this process, because of electric fields applied, charged particles accumulate at the nozzle, forming a conical meniscus known as a Taylor cone¹⁰⁶. Ultimately, the liquid is ejected from the cone’s tip onto the substrate¹⁰³. Sooman Lim et al. initially introduced EHD printing to the realm of 2D materials. They utilized ink based on 2D material powder for printing, leading to the production of incomplete 2D material films¹⁰³. In a subsequent study, Can et al. delved into the solution synthesis of MoS₂ precursor, and they applied the EHD jet printing technique to fashion patterned MoS₂ films, using (NH₄)₂MoS₄ solution as ink. This research resulted in the deposition of three layers of MoS₂ material¹⁰⁰. The width of the MoS₂ pattern is found to have an almost linear relationship with respect to stage speeds, with width of approximately 60 µm achievable at the highest speed. This method presents a promising solution for the precise printing of 2D materials, offering high resolution and the ability to control the number of layers effectively.

Aerosol Jet Printing (AJP) is also one of the methods to enhance the width resolution. AJP distinguishes itself as a non-contact, high-resolution printing approach that utilizes an atomizer to atomize ink into micron-scale droplets, thereby improving printing width resolution⁹³. In a study by Elahe Jabari et al., a graphene ink compatible with aerosol jet printing was developed using graphene powder as the raw material. It facilitated the printing of interconnects with controllable widths ranging from 10 to 90 µm on Si/SiO₂ substrates¹⁰¹; the setup is shown in Fig. 7e. The width control is achieved through the manipulation of the atomizer flow and the power of the atomization device. Similar methodologies have been employed in the aerosol jet printing of other 2D materials, including MoS₂¹⁰⁵, h-BN¹⁰⁴, etc. However, owing to the ink formulation, the outcome typically consists of lines containing flakes of 2D material rather than forming a continuous 2D material film. To address this limitation, the utilization of precursors convertible into 2D materials, such as ((NH₄)₂MoS₄), as inks hold promise, although this approach has not yet been published. Incorporating such precursors into the AJP process could potentially lead to the synthesis of continuous 2D material films, offering better width resolution compared to conventional methods.

Stamp printing offers a rapid method for obtaining patterned 2D material flakes through a simple process. However, its limited promotion is attributed to challenges in consistently controlling the thickness of 2D materials during printing. The inkjet printing method offers advantages such as room temperature and atmospheric pressure operation, ease of pattern realization using a printer, and minimal waste of raw materials. It is particularly useful for printing on flexible materials, making it suitable for applications in flexible electronics. Enhancements in printing resolution have been sought through emerging technologies like AJP (Aerosol Jet Printing) and EHD (Electrohydrodynamic) printing. However, post-printing annealing is often required to complete the conversion of 2D materials. Similar to the laser method, inkjet printing faces challenges in printing large areas and precisely controlling the thickness of the 2D material film. The formulation of the ink plays a crucial role and limits the range of synthesizable 2D materials using this method.

Atomic layer aligned stacking

Aligned stacking of 2D materials is an additive manufacturing method that involves stacking multiple layers of atomically thin materials in a specific orientation to form a 3D structure. In this process, the individual 2D material layers are bonded through covalent bonds, while the layers themselves are held together by van der Waals forces. Achieving a vertical heterostructure through manual assembly requires precise control and manipulation at the atomic level. By carefully aligning and stacking different 2D materials, researchers can create various properties and functionalities in the resulting structure.

In 1990, researchers at IBM moved individual xenon atoms on a nickel substrate, creating a patterned array with the help of scanning tunnelling microscopy (STM). It was a significant breakthrough, and demonstrate the potential for manipulating atoms¹⁰⁷. Afterward, atomic force microscopy (AFM) has also been utilised to manipulate single atoms and molecules due to its exceptional atomic-scale resolution^{108,109,110,111}. Scanning transmission electron microscopy (STEM) also proved to be used for imaging materials at the atomic scale but also for manipulating and modifying them. In the case of 2D materials, STEM has been used to dope the graphene to alter its electrical and chemical properties¹¹².

These atomic-scale explorations and tool development paved the way for aligned stacking at the atomic level. Yu et al. exfoliated MoS₂ layers were stacked on top of CVD-grown monolayer graphene using a micromechanical cleavage approach¹¹³. This technique enables the design of vertically stacked graphene-MoS₂-metal field-effect transistors (FETs). In the process described by Withers et al., a heterogeneous structure is created by stacking different kinds of 2D materials layer by layer¹¹⁴. The structure consists of dozens of layers and has an overall thickness of 10–40 atoms, one of the formed heterogeneous structures is shown in Fig. 8a, b. This is achieved through multiple “peel” and “lift” transfer processes using polymethyl methacrylate (PMMA). In each transfer step, a layer of the desired 2D material is peeled off from its original substrate using PMMA as a supporting layer and then lifted onto the growing stack of layers. However, it is important to note that these methods do not involve precise alignment, and the stacked layers are formed through a relatively simple stacking process.

Fig. 8: Aligned stacking for 2D materials.

a, b Schematic of the heterostructure hBN/GrB/2hBN/WS₂/2hBN/GrT/hBN and its cross-sectional bright-field STEM image. Reproduced from ref. 114 with permission from Springer Nature, Copyright 2015. c Optical microscope image of G/hBN heterostructure formed by robotic system. d Schematics of the vdW heterostructure fabrication process by robotic system. Reproduced from ref. 117 with permission from Springer Nature, Copyright 2018. e–h Formation of arbitrary 3D structures by the assembly of prepatterned 2D material layers: (e) two rotated crosses and (f) Hypothetical transistor structure consisting of different 2D material layers. g, h Schematic of the stacking process: putting patterned 2D material in the support frame; moving the membrane onto second support; the support frames moving laterally to rupture the membrane; letting 2D material stays on the target frame; repeating the step to get stacking layers. Reproduced from ref. 118 with permission from American Chemical Society, Copyright 2022.

To achieve high-quality transfers of 2D materials, an align transfer system has been developed, which includes both dry and wet methods¹¹⁵. Dry stacking is considered the mainstream approach in this context. The mechanical stacking platform, which is commonly used in dry stacking, features a typical structure comprising an optical microscopy-based imaging system, three-axis microscope stages, a rotary table for precise alignment, and a heating plate¹¹⁶. The transfer process based on the align transfer system exhibits a precision at the micron scale, ensuring accurate alignment and stacking of the 2D materials. However, it is important to address the issue of time-consuming manual operations and explore ways to improve efficiency. One promising solution is the use of robotic systems. Masubuchi et al. have designed a robotic system specifically for the automatic stacking of 2D materials¹¹⁷, the process is shown in Fig. 8d. This system significantly enhances efficiency and requires only a few minutes of manual participation per hour. It has the capability to identify and stack approximately 400 monolayer graphene flakes per hour, with a low error rate of less than 7%, the optical microscope image of the tacked structure is shown in Fig. 8c. The lateral error of the system is approximately 10 µm, providing precise positioning of the materials. Additionally, the operator can perform manual alignment, achieving a final alignment accuracy of less than ±1° and ±1 µm. Haas et al. have presented a method for creating 3D structures through the precise stacking of individual layers of 2D materials, as demonstrated in Fig. 8e, f¹¹⁸. The process involves several steps. Firstly, the structure of each individual 2D material layer, such as Tungsten disulfide (WS₂), Molybdenum disulfide (MoS₂), and graphene, are defined using a focused electron beam in TEM/STEM or electron-beam induced etching. Next, the aligned stacking of the layers is performed in a vacuum environment using a nanomanipulator, guided by visual feedback from SEM, as shown in Fig. 8g, h. The method achieves a remarkable precision in interlayer stacking, with an accuracy reaching 10 nm. It takes approximately 5 minutes to place one layer onto the stack. Based on this time estimation, the potential stacking speed can be estimated at around 10 nm3 per second.

Although atomic-level processing allows for precise control of atoms and atomic layers, its current throughput is limited, which may pose challenges for mass manufacturing. To enable widespread adoption and large-scale production, improvements are needed to enhance the efficiency and throughput of atomic-level processing techniques.

Applications

2D materials electronics

Over an extended period, integrated circuits reliant on silicon as the primary raw material have experienced rapid development in alignment with Moore’s law predictions. Nevertheless, it is evident that as silicon transistors approach their physical limitations, the expiration of Moore’s law looms. The pivotal role of 2D materials in extending Moore’s Law becomes apparent due to their ability to achieve single-atomic-layer thickness and exhibit ultra-high carrier mobility, demonstrating exceptional suitability as channel materials for field-effect transistors (FETs). The vertical stacking of 2D material FETs introduces the possibility of 3D transistor architectures^119,120. Despite there are lots of problems, such as doping and contact, that must be solved, the manufacturing phase has emerged as a critical bottleneck impeding its large-scale production.

The initial design of a graphene-based FET paved the way for further developments in the field¹²¹, the graphene films were prepared by mechanical exfoliation of mesas of highly oriented pyrolytic graphite, and transferred to the silicon dioxide, and were visually inspected to select appropriate graphene flakes for subsequent operations. Although mechanically exfoliated 2D materials are of high quality, the efficiency and consistency of this process are too low. TMDs have also shown great promise in next-generation FETs, thanks to their large on/off ratio, immunity to short-channel effects, and abrupt switching behavior¹²². Radisavljevic et al. designed monolayer MoS₂ transistors¹²³, The MoS₂-based FET performs better at room temperature, including mobility of 200 cm2 V⁻¹ s⁻¹, current on/off ratios of 1 × 10⁸ and ultralow standby power dissipation. Building upon the success of high-performance 2D material FETs, researchers have progressed towards integrated circuit (IC) design. In a study by the same group mentioned, an integrated circuit based on a monolayer of MoS₂ was developed to perform logic operations, specifically inverters, as depicted in Figure 9a¹¹⁹. Subsequently, Wang et al. expanded the use of MoS₂-based transistors to realize more complex circuits, including a NAND gate, a static random access memory (SRAM), and a five-stage ring oscillator¹²⁴. Despite numerous breakthroughs in the realm of 2D material devices, the fabrication method for these devices still relies on the original transfer process, utilizing mechanically exfoliated monolayers, akin to the initial graphene-based FET^119,123,124.

Fig. 9: 2D materials electronics and biosensors.

a Integrated circuit based on single-layer MoS₂. Reproduced from ref. 119 with permission from American Chemical Society, Copyright 2011. b Photograph of a MoS₂ based FET array on a 4 in. SiO₂/Si wafer fabrication by laser method. Reproduced from ref. 84 with permission from American Chemical Society, Copyright 2020. c Fiber SPR sensor with MoS₂-graphene structure. Reproduced from ref. 142 with permission from Elsevier, Copyright 2020. d Optical microscope image of a Graphene-based FET; (e) biosensing by the GFET. Reproduced from ref. 143 with permission from IOP Publishing, Copyright 2014.

In the subsequent phase of development, as efforts intensified to transition 2D material devices from laboratory research to industrial-scale applications, additional fabrication methods were introduced. Following the stabilization of the chemical vapor deposition (CVD) synthesis process for 2D materials, Sanne et al. presented a demonstration of MoS₂-based FETs. They utilized CVD-synthesized monolayer MoS₂ through the sulfurization of MoO₃. Electron beam lithography (EBL) and Cl₂ plasma etching were employed for the patterning process¹²⁵. MoS₂ FETs fabricated with the conventional gate-last process is also published¹²⁶. While these studies have significantly propelled the advancement of 2D materials by combining additive and subtractive manufacturing, the patterning process remains inefficient. In recent years, alternative additive manufacturing methods directly synthesizing 2D materials have emerged. One such approach involves the creation of transistors through the printing method, utilizing inkjet-printed graphene, MoS₂, and hBN¹²⁷. Additionally, MoS₂-based FETs have been successfully fabricated using laser-assisted methods^84,128. Through the laser-assisted technique, 1600 rectangular micro-patterns can be generated in just 5 minutes at a laser movement speed of 10 mm per second, forming a FET array, as depicted in Fig. 9b. This approach significantly enhances production efficiency. The advent of atomic-scale additive manufacturing technology is poised to play a pivotal role in future 2D material devices, facilitating the realization of mass production.

Biosensors

Due to their remarkable characteristics, including a high surface area to volume ratio and atomic thinness, 2D materials stand out among numerous nanomaterials, boasting the highest Surface Area per Gram (SAPG), single-layer graphene offers a theoretical maximum of 2630 m² per gram. The elevated SAPG values, coupled with low electrical noise, significantly enhance the capacity to detect analytes at low concentrations^129,130,131. Consequently, 2D materials are exceptionally well-suited for biosensing applications that demand heightened sensitivity, such as detecting proteins^132,133, nucleic acids^134,135,136 and whole cells^137,138,139 in vitro. Consequently, the development of additive manufacturing technologies for 2D materials, allowing for precise control over the number of layers, becomes indispensable for advancing biosensor applications. Biosensors based on 2D materials are mainly based on their properties in the electrical, optical and electrochemical fields. In the subsequent discussion, we will delve into the principles and production processes associated with these biosensors.

Electrochemistry sensors are mainly based on amperometry/voltammetry techniques and can convert the signal from the interaction between the receptor and the analyte into a measurable output. Generally, they involve a redox reaction at the working and reference electrodes¹⁴⁰. Sajid et al. designed an immune sensor based on printed MoS₂¹³², which realized the precise detection of PSA, IgG, and NF-κB by a large surface area, and the detection limits were 0.1 ng per ml, 1 ng per ml and 1 ng per ml. The sensor is generated through the printing of an ink comprising ultrafine crystalline powder of MoS₂, resulting in electrodes featuring a width of 50 µm and encompassing 20 to 40 finger pairs. This printing technique, inherently an additive manufacturing method, holds the promise of facilitating large-scale production. Optical sensors are based on the transmission and collection of light signals, including surface plasmon resonance (SPR) and fluorescence or Forster resonance energy transfer (FRET)¹⁴¹. Haixia Yu et al. designed an SPR sensor with MoS₂-graphene for glucose detection; the structure is shown in Figure 9c¹⁴². The MoS₂ and graphene are CVD synthesized. The study revealed that the sensitivity of the sensor varied with the number of layers of the 2D material, with the highest sensitivity achieved using three-layer MoS₂ and monolayer graphene. Consequently, there is a pressing need for a more controllable additive manufacturing process and devoid of patterns process, that allows for precise modulation of layer count. Such advancements are crucial for enhancing the sensor’s performance and enabling its mass production. Electrical sensors primarily utilize FET-based devices, where 2D materials can be employed as the gate electrode channel. In the case of biosensor FETs, the focus is on achieving high sensitivity to biomolecules rather than optimizing device performance metrics like speed and power consumption. The gate electrode’s exceptional sensitivity to the external environment enables the detection of conductivity changes in the channel resulting from biomolecule binding. This enables the conversion of biological signals into electrical signals. A typical structure of a Graphene-based FET is illustrated in Fig. 9d-e¹⁴³. The FET sensor can detect different analytes with different biological receptors bonding to the 2D material¹⁴⁴. Mannoor et al. showed the bio-selective detection of bacteria at the single-cell level by graphene FET through the self-assembly of antimicrobial peptides on the graphene surface¹³⁹. Zheng et al. designed graphene-based FET biosensors for DNA detection¹³⁶. The graphene layer was produced via CVD on a copper substrate and later transferred onto a SiO₂/Si substrate, which had been patterned using photolithography. Subsequently, peptide nucleic acid (PNA) was covalently immobilized on the graphene surface for functionalization.

Irrespective of the principles governing their operation, biosensors leverage the generous surface-to-volume ratio of 2D materials, highlighting the necessity for precise regulation of the thickness of the 2D material layer. Atomic-scale additive manufacturing methods offer superior control over the number of layers in 2D materials. Furthermore, contemporary biosensors, crafted through established processes, often confront challenges related to achieving high yield rates and efficiency. Implementing suitable additive manufacturing methods has the potential to simplify the manufacturing process, leading to improved efficiency and yield rates.

Nanoelectromechanical systems

Utilizing 2D materials enables the creation of atomically thin suspended film structures, enhancing the performance of Nanoelectromechanical Systems (NEMS) and opening up possibilities in the field. Nevertheless, certain manufacturing challenges associated with 2D materials inevitably constrain their development and application within the field. Despite these limitations, there are numerous applications of 2D materials in the realm of NEMS, including pressure sensors^145,146,147, microphones^148,149, gas sensor^150,151, etc.

Based on the piezoresistive effect induced by mechanical strain in the graphene, Smith et al. designed piezoresistive pressure sensors with suspended graphene membranes¹⁴⁵. The same kind of sensor is also made 4.5 nm thick PtSe₂ with the similar process by Wagner et al., shown in Fig. 10a¹⁴⁶, the average sensitivity is 5.51 × 10 ^–4 mbar ^–1. The thickness of the suspended membrane greatly influences the sensitivity of the electromechanical transducer. Both devices are manufactured through the transfer of a 2D material flake. While these devices have exhibited commendable performance, achieving batch manufacturing can pose challenges. Certain additive manufacturing methods, as discussed earlier, such as stamp printing, can print 2D materials array in a single transfer step, which may be helpful for the high-volume production of devices.

Fig. 10: Nanoelectromechanical systems based on 2D materials.

a NEMS piezoresistive pressure sensors with Platinum diselenide (PtSe2) membranes. Reproduced from ref. 146 with permission from American Chemical Society, Copyright 2018. b photograph of the graphene membrane in the microphone holder(left) and the internal structure(right). Reproduced from ref. 148 with permission from IOP Publishing, Copyright 2014.

Lee et al. designed MoS₂ resonators and compared the effect of the number of MoS₂ layers on the performance of the resonator, the result shows that mono-, bi-, and tri-layer MoS₂ devices have lower FMF damping than thicker, conventional devices¹⁴⁷. This underscores the ability of few layers of 2D materials to enhance the properties of NEMS devices. The MoS₂ layers in the resonators were acquired through mechanical stripping, as discussed earlier. Graphene is also used in microphones and ultrasonic detection. Todorović et al. designed a kind of condenser microphone based on 25 nm multilayer graphene, as shown in Fig. 10b. The sensitivity of it is improved 12 dB, at frequencies up to 11 kHz¹⁴⁸. Wittmann et al. endeavored to create graphene microphones through a CMOS-compatible process, achieving success. However, in practical applications, the overall efficiency is notably low due to the use of a wet transfer technique for transferring graphene sheets¹⁴⁹, which aligns with the challenges encountered by other 2D materials NEMS devices. Thanks to the large surface area to volume ratio of 2D materials, NEMS systems based on 2D materials are also used in environmental monitoring, such as air monitoring. Yang et al. designed graphene-based flexible NO₂ sensors on paper substrates¹⁵¹. The sensor was exposed to 200ppm NO₂ gas and would immediately generate an electrical signal response. The graphene film used in this sensor is generated and transferred by CVD. Sensors crafted on a flexible substrate are well-matched for specific additive manufacturing techniques, notably laser-assisted methods and inkjet printing. These approaches prove to be considerably more efficient compared to post-synthesis transfer methods involving CVD. Vertically Aligned MoS₂ Layers are used to enhance the gas adsorption capacity by Cho et al.¹⁵⁰. Compared with the horizontally arranged MoS₂ Layers, the detection ability of the former to the target gas NO₂ is increased by 5 times. The process involves obtaining MoS₂ through CVD sulfide Mo seed layer. This transfer-free method, enabling direct CVD synthesis of desired 2D materials, holds potential applicability in various other applications.

While 2D materials present numerous possibilities in NEMS, their development and commercial application have been hampered by process limitations. Currently, most NEMS devices relying on 2D materials are acquired through transfer processes. Despite the development of NEMS processes compatible with traditional CMOS processes, the transfer of 2D materials often leads to reduced efficiency. To address these challenges, additive manufacturing methods emerge as potential solutions. For instance, stamp printing methods enable the controlled transfer of multiple pieces of 2D materials simultaneously. Printing and laser techniques can generate 2D materials directly on flexible substrates. Additionally, certain direct CVD synthesis methods have the potential to significantly enhance efficiency.

Energy conversion and storage

The substantial specific surfaces of 2D materials enhance ion adsorption, leading to increased capacitance. Their high conductivity accelerates electron transport, and the presence of adjustable active sites enables electrocatalytic activity¹⁵². These advantages render 2D materials versatile for various applications in energy conversion and storage.

Supercapacitors are a focal point in contemporary energy storage research. Huang et al. employed dispersed graphene derived from few-layer graphene sheets exfoliated through electrochemical methods as the primary material to fabricate a graphene/cellulose (GC) hybrid film for use as a supercapacitor electrode¹⁵³. However, this process involves numerous intricate steps and presents challenges for large-scale preparation. Wang et al. synthesized 1T-VSe2 nanosheets, a Transition Metal Dichalcogenide (TMDs) material, using Chemical Vapor Deposition (CVD). Subsequently, flexible supercapacitors were created by transferring these nanosheets as electrode materials¹⁵⁴. Although the CVD method provides greater control compared to the stripping method, the overall efficiency of the combination of CVD and transfer is relatively lower. If laser-assisted or printing methods are employed for direct additive manufacturing on flexible substrates, production efficiency could be significantly enhanced. 2D materials find applications in metal-ion batteries as well. Hu et al. utilized exfoliated TiS₂ nanosheets, TMDs, in Sodium-Ion Batteries¹⁵⁵. Similarly, Sun et al. developed MoS₂/graphene nanosheets by ball milling and exfoliation of bulky MoS₂ and graphite, serving as anode materials for high-rate sodium-ion Batteries¹⁵⁶. Both approaches hold the potential to enhance efficiency through the incorporation of 2D materials in printing inks. Alinejadian et al. fabricated a MoS₂/Mo₂S₃ nanocomposite with promising potential as an anode material for future storage devices due to its favorable electrical activity^94,95. The composite is manufactured using the additive manufacturing method known as selective laser melting (SLM) due to its high efficiency and flexibility.

The application of 2D materials in energy storage is extensive, and a prevailing challenge in the field is enhancing production efficiency. Certain additive manufacturing methods, such as laser-assisted techniques and printing, hold the potential to address this challenge and improve overall efficiency. Moreover, practices grounded in laser methods have demonstrated the effectiveness of additive manufacturing in advancing efficiency within this domain.

Discussion and future perspectives

Atomic-scale additive manufacturing of 2D materials is essential to meet the requirements of atomic-level devices. The precise atomic layer control achieved through additive manufacturing ensures uniformity and consistency in material thickness and composition. This is crucial for achieving the desired device properties, such as electrical conductivity, mechanical strength, and sensitivity. The capability of controlling the atomic layer thickness allows for device miniaturization, enabling the fabrication of smaller and more compact devices with improved performance. Furthermore, the precise positioning control offered by atomic-scale additive manufacturing optimizes the manufacturing process and reduces the reliance on traditional photolithography techniques. This eliminates the need for complex lithographic masks and reduces the introduction of impurities, improving manufacturing efficiency and scalability.

Various potential atomic-scale additive manufacturing methods were discussed in this article. Based on these methods, the goal is to achieve precise control of the number of 2D material layers, precise control of the generation area, increase manufacturing speed, reduce process complexity, and cover more types of 2D materials. Based on these factors, the mentioned methods are evaluated and the result is shown in Table 7. Mainstream methods like CVD and ALD offer significant advantages in terms of the quality and coverage of synthesized 2D materials. However, these methods require high-end equipment and stringent environmental conditions, making them less accessible and challenging to achieve patterning without lithography. Alternative methods such as electrodeposition, laser-assisted synthesis, and printing techniques offer the advantage of easier patterning without the need for expensive equipment or strict environmental conditions. However, these methods often suffer from limitations in terms of the quality and coverage of the synthesized 2D materials. Achieving high-quality and uniform 2D material synthesis remains a challenge for these approaches. Another approach, known as aligned stacking of 2D materials, allows for precise control over the thickness of the resulting 3D structure and the formation of heterojunctions by stacking different materials. However, this method typically involves slow speeds and heavily relies on manual operation, limiting its scalability and efficiency. Overall, each of these methods has its own strengths and limitations when it comes to the additive manufacturing of 2D materials.

Table 7 Summary of the characteristics of various 2D material additive manufacturing methods

While there have been significant advancements in atomic-scale additive manufacturing of 2D materials, several challenges still need to be addressed. One key challenge is achieving selective synthesis of 2D materials with precise control over dimensions, particularly achieving line widths down to a few nanometers. Currently, the accuracy of existing additive manufacturing methods is limited to a few microns. However, this level of precision would not meet the stringent requirements for some applications such as electronic devices. A photolithography-free manufacturing process can only be realized by attaining genuine high resolution across the width. Creating patterns without photolithography poses challenges for methods such as CVD, ALD, and electrodeposition, and the use of plasma to treat substrates emerges as a promising approach in overcoming these challenges. Printing methods offer potential improvements in resolution through innovative technologies for controlling ink droplets. Among these methods, laser-based techniques are the most likely to achieve breakthroughs in width resolution, accomplished through the modulation of laser bands or pulses. Additionally, achieving accurate monolayer synthesis of 2D materials in selected areas is another important area of exploration. This would enable precise control of 2D material thickness, allowing for optimization of device performance. This bottleneck primarily affects electrodeposition methods, laser methods, and printing methods. Overcoming this challenge necessitates the optimization of various parameters, including precursor selection, precursor concentration, processing time, and other proprietary factors for each method. Lastly, it is crucial to develop a stable and reliable process to enable large-scale manufacturing and widespread application of atomic-scale additive manufacturing of 2D materials. This process should ensure consistent quality of the generated 2D materials and a high yield of atomic-level devices. Meeting this requirement presents challenges for all the additive manufacturing methods mentioned, particularly in the stacking of individual atomic layers. While this method allows for precise control over the number of 2D material layers, it currently faces challenges in terms of low operating efficiency and inconsistent formation of devices. These challenges may be solved by AI-based robotic systems. Addressing consistency issues is also imperative for laser methods, printing methods, and electrodeposition methods; it also requires optimization of various parameters. Only a method capable of simultaneously addressing all three challenges stands a chance of realizing true atomic-scale additive manufacturing.

The successful advancement of atomic-scale additive manufacturing for 2D materials holds significant potential across various applications. In electronics, the highly efficient and precise additive manufacturing process for 2D material components is poised to revolutionize the mass production of electronic devices. Additionally, additive manufacturing methods contribute to achieving vertical stacking of transistors, forming 3D structures that maximize space utilization. This advancement becomes particularly crucial for enhancing transistor density and computational capabilities beyond the limitations of Moore’s Law. Within the biosensing field, atomic-scale additive manufacturing shows immense promise in enhancing sensor yield and quality. Sensors produced at this scale can detect and analyze biomarkers with unparalleled sensitivity, leading to breakthroughs in disease diagnosis and biomedical research. Stable manufacturing techniques also contribute to improving the production yield and efficiency of biosensors, promoting the commercialization of highly sensitive sensors for the benefit of a larger population. In the realm of NEMS, atomic-scale additive manufacturing plays a pivotal role. This technology enables the creation of intricate structures with atomic precision, resulting in sensors with advanced functionalities. These sensors can be seamlessly integrated into diverse devices and systems, propelling the evolution of the Internet of Things (IoT) and fostering a more interconnected and intelligent world. Within energy conversion and storage, additive manufacturing methods extend their utility by facilitating the broader application and more efficient production of 2D materials. This progress contributes significantly to the advancement of energy storage and renewable energy technologies.

This review article underscores the potential and significance of atomic-scale additive manufacturing in the fabrication of 2D materials. Through a comprehensive analysis of various methods, including site-selective chemical vapor deposition, area-selective atomic layer deposition, electrodeposition, laser-assisted synthesis, print method, and atomic layer-aligned stacking, the advantages and disadvantages of each technique are meticulously evaluated. The systematic discussion of these methods provides valuable insights for driving future advancements in manufacturing technology. Considering the specific application requirements of atomic-scale additive manufacturing for 2D materials, the review identifies key areas for future research. The refinement of selective synthesis methods with increased precision and reduced scale emerges as a priority. Meanwhile, the development of accurate monolayer synthesis techniques becomes crucial in attaining precise control over the fabrication process. Once the precision of synthesis is achieved, the establishment of stable and reliable processes for large-scale manufacturing becomes essential to promote the widespread adoption of additive manufacturing at the atomic and close-to-atomic scale. In conclusion, this review offers a profound comprehension of atomic-scale additive manufacturing for 2D materials and provides a solid foundation for forthcoming breakthroughs in this dynamic and rapidly evolving area.