Commercial capacitive pressure sensors often lack sensitivity. This research addresses this limitation by creating a sensor array with a large number of parallel-connected membranes. The use of few-atom-thick DLG membranes allows for competitive performance with commercial alternatives. The goal is to design a low-cost, portable, battery-powered sensor for pressure detection, leveraging the unique mechanical and electrical properties of graphene. The study's importance lies in the potential for widespread applications requiring sensitive and affordable pressure measurement.

Previous research demonstrated that arrays of graphene drums are needed to achieve commercially competitive sensitivity. Studies have also explored the use of protective polymer layers like PMMA to improve the yield and mechanical strength of graphene-based devices. However, these polymer layers can reduce membrane deflection and sensor responsivity. This work builds upon these findings by investigating methods to improve the sensor's performance by optimizing the design and removing the limiting polymer layer.

The sensor chip design features approximately 10,000 circular graphene drums (5 µm diameter) arranged on a hexagonal grid (10 µm pitch) on a 1 × 1 mm² silicon chip with a 285 nm SiO₂ layer. Ti/Au electrodes (5 nm/60 nm) are patterned for contacting the graphene top electrode. Circular holes (240 nm deep) are etched into the SiO₂ to suspend the DLG membranes, preventing short circuits. Double-layer graphene is synthesized via CVD and transferred onto the chip using PMMA as a support layer (800 nm thick). A yield of 95–99% freely suspended DLG/PMMA membranes is achieved. The readout circuit uses a commercial capacitance-to-digital converter (AMS PCap04) to measure capacitance changes caused by pressure-induced membrane deflection. An Arduino Uno processes the data, displaying pressure on an LCD. To improve the sensor's performance, the PMMA layer is removed through thermal annealing (300 °C, 500 Torr, 0.5 SLPM Ar or N₂). The annealing process reduces PMMA thickness at a rate of ~27 nm/min. AFM and Raman spectroscopy characterize the graphene membranes before and after annealing. Force-indentation AFM measures the mechanical properties (Young's modulus) of the membranes. Equation (1) defines the noise floor (NF), and equation (2) models the point deflection of a circular membrane. Equation (3) relates pressure difference to membrane deflection.

The thermal annealing process significantly enhances sensor responsivity, nearly two orders of magnitude in the best case. The noise floor also decreases substantially after PMMA removal. AFM imaging reveals three drum states: intact, ruptured, and collapsed. Only intact drums contribute significantly to the pressure response. Post-annealing AFM shows PMMA residue accumulation, mostly between the graphene layers. Force-indentation AFM measurements yield a mean two-dimensional Young's modulus of 175 N/m, lower than reported values for pristine single-layer graphene but comparable to other CVD graphene membranes and 2D materials. Raman spectroscopy shows that the DLG remains largely defect-free after annealing, with minor changes in the 2D and G peak positions and ratios attributed to variations in the twist angle and strain. The sensor exhibits a linear response at small deflections, transitioning to nonlinear behavior at larger deflections due to mechanical and capacitance effects.

The significant improvement in responsivity and noise floor after PMMA removal demonstrates the effectiveness of the thermal annealing process. The lower Young's modulus compared to pristine graphene likely results from the presence of PMMA residues between the graphene layers, affecting the membrane's mechanical properties. The observed linearity at small deflections is consistent with the theoretical model, while the nonlinearity at larger deflections suggests the need for more refined modeling to capture the complex interplay between mechanical and electrical effects. The high yield and relatively low defect density after annealing highlight the robustness of the fabrication process and the resilience of DLG to high-temperature processing. These findings suggest promising avenues for the development of high-performance, low-cost pressure sensors for various applications.

This research successfully demonstrates a highly sensitive pressure sensor based on an array of DLG membranes. The optimized design, coupled with the thermal annealing process for PMMA removal, yields a significant improvement in responsivity and noise floor. Future work could focus on further optimizing the chip design, exploring alternative materials for improved performance, and investigating applications in various fields. Further refinement of the model to accurately predict behavior across larger deflections would improve the sensor's calibration and accuracy.

The study primarily focuses on a specific fabrication process and material combination. The generalization to other graphene qualities or fabrication methods needs further investigation. The presence of residual PMMA between the graphene layers may affect long-term stability and performance. The current design's sensitivity might be affected by temperature variations and requires further characterization for different temperature ranges.