Low-temperature growth of 2D materials is critical for back-end-of-line (BEOL) integration in semiconductor manufacturing, where thermal budgets are constrained to below 200°C to avoid damaging pre-existing metal interconnects and dielectric layers. Traditional high-temperature synthesis methods, such as chemical vapor deposition (CVD) above 600°C, are incompatible with BEOL processes. Instead, alternative techniques have been developed to enable direct deposition of 2D materials on CMOS-compatible substrates without post-growth annealing. These methods must address key challenges, including precursor reactivity limitations and achieving sufficient crystallinity at reduced temperatures.

Plasma-enhanced chemical vapor deposition (PECVD) is a prominent low-temperature technique for 2D material growth. By using plasma to dissociate precursor gases, PECVD lowers the activation energy required for chemical reactions, enabling deposition at temperatures as low as 100°C. For example, graphene-like films have been synthesized using methane and hydrogen precursors under RF plasma excitation. However, PECVD-grown materials often exhibit higher defect densities compared to high-temperature CVD due to incomplete precursor decomposition and insufficient adatom mobility. The crystallinity of PECVD-grown transition metal dichalcogenides (TMDCs), such as MoS2, is typically lower, with domain sizes limited to tens of nanometers. Despite this, PECVD remains attractive for BEOL integration due to its scalability and compatibility with industrial tools.

Atomic layer deposition (ALD) is another low-temperature method capable of producing uniform, conformal 2D films. ALD relies on self-limiting surface reactions between alternating precursor pulses, allowing precise thickness control at the atomic level. For TMDCs like WS2, metal-organic precursors such as tungsten hexacarbonyl and hydrogen sulfide have been used at temperatures between 150°C and 200°C. The primary challenge with ALD is achieving stoichiometric control, as low temperatures can lead to incomplete ligand exchange or carbon contamination from precursor residues. Post-deposition treatments are generally avoided in BEOL-compatible processes, necessitating careful precursor selection and process optimization to minimize impurities.

Solution-based methods, including liquid-phase exfoliation and electrochemical deposition, offer room-temperature synthesis options. Liquid-phase exfoliation involves dispersing bulk layered materials in solvents or surfactants to produce 2D flakes, which can then be deposited via spin-coating or inkjet printing. While this approach avoids high temperatures altogether, the resulting films are typically polycrystalline with random flake orientation, leading to poor electronic properties. Electrochemical deposition, on the other hand, can produce more continuous films by reducing metal ions and chalcogens directly on substrates. For instance, MoS2 has been electrochemically deposited at temperatures below 100°C using ammonium tetrathiomolybdate as a precursor. However, controlling film stoichiometry and minimizing oxygen incorporation remain significant hurdles.

Metal-organic chemical vapor deposition (MOCVD) at low temperatures has also been explored for 2D material growth. By using highly reactive metal-organic precursors, such as molybdenum hexacarbonyl and diethyl sulfide for MoS2, deposition can occur at temperatures around 150°C. The trade-off is increased carbon contamination due to incomplete precursor decomposition, which degrades electrical performance. Additionally, the high vapor pressure of some metal-organic precursors complicates process control, requiring precise flow and pressure regulation to maintain uniformity.

Precursor reactivity is a central challenge in low-temperature 2D material growth. At reduced temperatures, conventional precursors like elemental sulfur or selenium exhibit insufficient reactivity with metal sources, leading to incomplete chalcogen incorporation. To mitigate this, alternative precursors such as hydrogen sulfide or alkyl chalcogenides are employed, but these can introduce unwanted impurities or require additional purification steps. For example, using tert-butyl disulfide as a sulfur precursor in MoS2 growth improves reactivity at 200°C but may leave residual carbon in the film. Similarly, metal-organic precursors for TMDCs often contain ligands that are difficult to fully remove at low temperatures, affecting film quality.

Crystallinity is another major limitation in low-temperature synthesis. The reduced thermal energy available at sub-200°C conditions restricts adatom surface diffusion, resulting in smaller grain sizes and higher defect densities. In the case of graphene, low-temperature PECVD films typically exhibit a high density of grain boundaries and sp3-hybridized carbon defects, leading to lower carrier mobility compared to high-temperature CVD graphene. For TMDCs, the domain size of low-temperature films is often below 100 nm, which can degrade optoelectronic performance due to enhanced scattering at grain boundaries. Strategies to improve crystallinity without increasing temperature include plasma-assisted processes, where ion bombardment can enhance adatom mobility, or the use of catalytic substrates that lower the energy barrier for crystallization.

BEOL integration imposes additional constraints beyond temperature, including compatibility with dielectric layers and metal interconnects. For instance, growth on silicon dioxide or low-k dielectrics requires careful control of nucleation density to prevent excessive film roughness or delamination. Direct growth on copper or aluminum interconnects is particularly challenging due to potential metal diffusion or interfacial reactions at low temperatures. Some approaches employ sacrificial buffer layers or seed layers to promote uniform 2D material growth while protecting underlying BEOL structures. For example, a thin alumina layer can serve as a diffusion barrier for metal interconnects during MoS2 deposition.

Applications of low-temperature 2D materials in BEOL integration include interconnects, transistors, and sensors. Graphene and TMDCs have been explored as potential replacements for copper interconnects in advanced nodes, where their high carrier mobility and atomic thickness could mitigate resistance-capacitance delays. However, the defect-limited mobility of low-temperature films remains a bottleneck. In transistors, BEOL-compatible MoS2 FETs have been demonstrated with on/off ratios exceeding 10^6, but their field-effect mobility is typically below 10 cm^2/Vs due to grain boundary scattering. For sensors, the large surface-to-volume ratio of 2D materials enables high sensitivity in gas or biosensing applications, where low-temperature growth allows integration directly atop CMOS circuits.

Future advancements in low-temperature 2D material growth will likely focus on precursor engineering and process optimization. Developing novel precursors with higher reactivity at reduced temperatures could improve film quality, while advanced plasma or photonic activation methods may enhance crystallinity without exceeding thermal budgets. Additionally, machine learning-driven process optimization could help identify ideal growth conditions for specific BEOL applications, balancing trade-offs between crystallinity, purity, and throughput. As the demand for monolithic 3D integration grows, the ability to deposit high-performance 2D materials at BEOL-compatible temperatures will become increasingly critical for next-generation semiconductor technologies.