Nonlocal Electrostatic Origin of Schottky-Barrier Variability in 2D Contacts

Technical Summary: Nonlocal Electrostatic Origin of Schottky-Barrier Variability in 2D Contacts

Problem Statement
Electrical contacts frequently limit the performance of atomically thin semiconductor devices. A persistent challenge in the field is the wide variability in reported Schottky barrier height (SBH) values for identical metal/2D-semiconductor pairs. For instance, reported SBHs for Au–MoS2 contacts range from 0.05 to 0.95 eV, and for Ti–MoS2 from 0.05 to 0.46 eV. Conventional theoretical frameworks treat SBH as a local interface property, attributing deviations from the Schottky–Mott limit to interface gap states and Fermi-level pinning. In this view, microscopic disorder is expected to be averaged out by interfacial charge reservoirs or constrained by the "pinch-off" effect of surrounding high-barrier regions. However, the experimentally observed "giant spread" in SBH values cannot be reconciled with a model where the barrier is determined solely by local interface properties.

Methodology
The authors propose a nonlocal electrostatic model where the effective SBH in top-contacted 2D devices is treated as an edge observable driven by remote electrostatic perturbations, rather than a spatial distribution of local barriers.

1. Theoretical Framework:

   • The study considers a top-contact geometry with a covered region ($x < 0$), an uncovered channel ($x > 0$), and a contact edge at $x = 0$.
   • The lateral SBH ($\Phi^{\text{lat}}_B$) is defined as the conduction-band step across the edge. The model posits that the barrier shift is controlled by the change in the edge lineup ($\Delta J$), which is linearly coupled to the electrostatic potential at the edge, $u(0)$.
   • Using a quadratic free-energy functional for the electrostatic potential, the authors derive a response kernel that maps a remote defect source $s_d(x)$ to the edge potential.
   • The resulting barrier shift due to a point defect at distance $x_d$ is given by:
     $$\Delta \Phi^{\text{lat}}_B (x_d) = \chi_e S_d \frac{\lambda_l}{\kappa + K_e \lambda_l} e^{-x_d/\lambda_l}$$
     where $\lambda_l$ is the propagation length determined by the channel's in-plane electrostatic stiffness ($\kappa$) and restoring capacitance ($C_{ch}$), and the amplitude is determined by contact-specific parameters (edge susceptibility $\chi_e$ and edge stiffness $K_e$).

2. Computational Validation:

   • The model is validated against first-principles calculations using Density Functional Theory (DFT) combined with the Nonequilibrium Green's Function (NEGF) method.
   • Systems studied include Ti–MoS2 and Au–MoS2 top contacts.
   • Sulfur vacancies are introduced in the MoS2 channel at varying distances ($0d$ to $10d$) from the contact edge.
   • Effective SBHs are extracted from the local density of states (LDOS) near the contact edge, specifically measuring the conduction-band step.

Key Contributions and Results

• Nonlocal Mechanism: The paper establishes that defects located away from the contact edge can significantly modulate the measured SBH through nonlocal electrostatic coupling. The influence of a defect decays exponentially with distance, governed by a characteristic length scale $\lambda_l$.
• Decoupling of Roles: The model reveals a clear division of roles:
  • The semiconductor channel determines the range of the nonlocal effect ($\lambda_l$), which depends on in-plane electrostatic spreading and screening.
  • The metal contact determines the amplitude of the SBH shift, governed by the edge's susceptibility to potential changes and the efficiency of converting edge potential shifts into lineup modulation.
• Quantitative Agreement:
  • DFT-NEGF calculations for both Ti–MoS2 and Au–MoS2 show that the lateral SBH follows the predicted exponential decay with defect distance.
  • A common interaction length of $\lambda_l \approx 1.1$ nm fits the data for both metals, confirming that this length scale is primarily a property of the MoS2 channel electrostatics rather than the specific metal.
  • The calculated SBH variations span a range that accounts for a substantial portion of the experimentally reported variability for Ti–MoS2 and a significant part of the spread for Au–MoS2.
• Unified Explanation: The results provide a unified explanation for the long-standing variability in extracted SBHs, demonstrating that nominally similar contacts can yield vastly different barrier heights depending on the spatial distribution of defects relative to the contact edge.

Significance
The paper claims to resolve a major inconsistency in the 2D-contact literature by demonstrating that the measured Schottky barrier is not a purely local interface property but a nonlocal, edge-controlled quantity. By identifying the edge environment and the spatial distribution of nearby defects as key design variables, the work suggests that carrier injection in 2D devices is governed by mesoscopic disorder environments rather than just atomic-scale interface structures. This framework moves beyond classical inhomogeneous-Schottky and pinch-off pictures, offering a mechanism for the "giant spread" in SBH values and establishing nonlocal electrostatics mediated by edge-proximal defects as a fundamental factor in 2D contact performance.