2D materials CMOS integration is being explored as a way to extend the capabilities of silicon-based systems, not simply replace them. Atomically thin semiconductors could offer useful electrostatic control at small dimensions and enable device concepts for low-power electronics. For industrial automation, these developments are relevant to the longer-term design of efficient processors, connected devices, and integrated sensing or computing platforms. But the gap between promising device demonstrations and manufacturable technology remains substantial.
Why 2D materials are attracting attention
Unlike conventional bulk crystals, 2D materials consist of strongly bonded atomic layers held together by weaker interactions between layers. Their surfaces have no dangling bonds, and individual layers are only a fraction of a nanometre thick. This structure can provide an atomically thin transistor channel with a clean interface, while limiting some short-channel effects that complicate the scaling of silicon devices.
Among the materials under study, transition metal dichalcogenides (TMDCs) such as molybdenum disulfide (MoS2) and tungsten diselenide (WSe2) are of particular interest because they are semiconductors with a bandgap. That property supports transistor operation with a distinction between ON and OFF states. Graphene, by contrast, is highly conductive but semi-metallic, which limits its suitability as the channel material in conventional digital logic. It is more often considered for roles such as interconnects.
These characteristics make 2D materials potential companions to CMOS for specific functions. They may also enable new transistor architectures, although the reported advantages of simulated devices have not yet been consistently matched by experimental hardware.
Manufacturing integration starts with the material
Mechanical exfoliation—the removal of thin flakes from a layered crystal—has produced high-quality samples that help researchers investigate material properties and demonstrate devices. However, flakes vary in thickness, size, and shape, making this approach unsuitable for scalable manufacturing.
Two broader strategies are being investigated:
- Direct growth: The material is synthesized on the wafer or device substrate where it will be used. This avoids a transfer step, but growth temperature and process chemistry must be compatible with the underlying CMOS structure.
- Growth followed by transfer: The material is grown on a substrate selected for better synthesis conditions, then moved to the target wafer. This separates growth requirements from device integration but adds a process step that can introduce contamination or damage.
The temperature constraint is particularly important for back-end-of-the-line (BEOL) integration, where existing circuitry must remain within its process limits. The reviewed work identifies temperatures around 450°C as a challenge for high-quality TMDC growth: material produced under such constraints can be highly defective. Transfer offers more flexibility in growth conditions, but repeatable, large-area transfer without altering the material remains difficult.
Because a 2D film is so thin, the receiving surface matters. Roughness, topography, surface chemistry, and contamination can affect the material, potentially causing unintended doping, strain, or defects. Those changes can in turn influence transistor stability and electrical behavior. Integration therefore requires control not only of the 2D layer but also of the substrate and interfaces around it.
Heterostructures and emerging device options
Combining different 2D materials into vertical stacks or lateral junctions creates heterostructures. These arrangements are being explored for devices such as tunnel field-effect transistors (TFETs), which use band-to-band tunneling rather than the drift-diffusion mechanism of conventional MOSFETs. Atomically abrupt interfaces are important to their operation. Growing a heterostructure in place may reduce interface contamination compared with assembling separate layers by transfer, although both approaches have been demonstrated.
TFETs and negative-capacitance FETs (NC-FETs) are two device concepts under investigation for low-power operation. TFETs aim to achieve steep switching through tunneling. NC-FETs use a ferroelectric gate material to influence the channel response. Experimental 2D-device demonstrations show potential, but issues such as interface traps, contacts, defects, and—in NC-FETs—hysteresis continue to affect device behavior. A demonstration of steep switching alone does not establish suitability for a complete, reliable circuit.
Controlling transistor polarity without conventional doping
Conventional ion implantation does not translate readily to atomically thin channels, and chemical doping methods can face challenges with stability and CMOS compatibility. An alternative is a polarity-controllable transistor that uses separate gates: a control gate to switch the channel and a polarity gate to influence carrier injection at the contacts. By changing the polarity-gate bias, an ambipolar semiconductor can be configured for n-type or p-type operation without fixed physical doping.
WSe2 devices have demonstrated this approach in logic gates including NAND, NOR, and XOR. The concept could allow more flexible circuit design, but its circuit demonstrations do not resolve the broader manufacturing requirements for wafer-scale uniformity, dependable interfaces, or high-volume process control.
What must be solved before broader adoption
For 2D devices to move toward CMOS-compatible systems, several requirements must advance together:
- Large-area synthesis with controlled thickness and low defect density.
- Uniform, repeatable transfer or direct growth at suitable process temperatures.
- Low-resistance, reliable contacts and controlled electrical polarity.
- Characterization methods that can assess material quality and interfaces across wafers.
- Device and circuit models that account for contacts, defects, and other non-idealities.
These are linked challenges. For example, a device architecture may depend on a particular interface, while manufacturing yield depends on reproducing that interface consistently. Material selection also needs to reflect the intended application rather than treating all 2D materials as interchangeable.
Practical implications for automation decision-makers
For plant and systems teams, 2D electronics should currently be understood as an emerging research direction, not a drop-in replacement for established silicon platforms. Potential relevance lies in future low-power processors and other electronics supporting connected or energy-constrained systems, but the cited work does not establish production readiness or quantify operational benefits for automation equipment.
The practical takeaway is to track progress in integration as closely as device performance. Wafer-scale quality, process compatibility, stable interfaces, and repeatable electrical characteristics will determine whether promising 2D transistor concepts can become useful components in larger systems. Until those manufacturing milestones are met, the strongest value for technical decision-makers is a clear understanding of where the technology may fit—and which engineering barriers still stand in the way.