Graphene-aluminum nitride NEMS resonant infrared detector

Abstract

The use of micro-/nanoelectromechanical resonators for the room temperature detection of electromagnetic radiation at infrared frequencies has recently been investigated, showing thermal detection capabilities that could potentially outperform conventional microbolometers. The scaling of the device thickness in the nanometer range and the achievement of high infrared absorption in such a subwavelength thickness, without sacrificing the electromechanical performance, are the two key challenges for the implementation of fast, high-resolution micro-/nanoelectromechanical resonant infrared detectors. In this paper, we show that by using a virtually massless, high-electrical-conductivity, and transparent graphene electrode, floating at the van der Waals separation of a few angstroms from a piezoelectric aluminum nitride nanoplate, it is possible to implement ultrathin (460 nm) piezoelectric nanomechanical resonant structures with improved electromechanical performance (>50% improved frequency×quality factor) and infrared detection capabilities (>100× improved infrared absorptance) compared with metal-electrode counterparts, despite their reduced volumes. The intrinsic infrared absorption capabilities of a submicron thin graphene-aluminum nitride plate backed with a metal electrode are investigated for the first time and exploited for the first experimental demonstration of a piezoelectric nanoelectromechanical resonant thermal detector with enhanced infrared absorptance in a reduced volume. Moreover, the combination of electromagnetic and piezoelectric resonances provided by the same graphene-aluminum nitride-metal stack allows the proposed device to selectively detect short-wavelength infrared radiation (by tailoring the thickness of aluminum nitride) with unprecedented electromechanical performance and thermal capabilities. These attributes potentially lead to the development of uncooled infrared detectors suitable for the implementation of high performance, miniaturized and power-efficient multispectral infrared imaging systems.

Keywords: MEMS; NEMS; aluminum nitride; graphene; infrared detector; piezoelectric; resonant sensor.

Figure 1

Graphene electrode for NEMS resonant IR detectors. (a) Schematic illustration in layer view of a G–AlN NEMS resonator vibrating in its contour-extensional mode. The top electrode, which is critical in confining the electric field within the AlN nanoplate membrane, is fabricated using mechanically transferred, CVD-synthesized graphene. The color map displayed on the AlN layer is created by 3D FEM simulation, which indicates the locations of maximum (red) and minimum (blue) mechanical displacement of the contour-extensional mode in the AlN nanoplate. (b) A false-colored tilted-SEM image of a fabricated G–AlN NEMS resonant IR detector. The device is 75-μm wide and 200-μm long.

Figure 2

Microfabrication process of G–AlN and reference devices. (a) Mask 1—sputter deposition and lift-off of Pt bottom electrode; (b) Mask 2—sputter deposition of AlN, wet etch to open vias; (c) Mask 3—dry etch to define device lateral dimensions; (d) Mask 4—sputter deposition and lift-off of top Au probing pads; (e) Mask 5—top electrode synthesis (graphene transfer or gold deposition) and patterning; (f) XeF₂ dry release of the AlN resonators (no mask required). Steps a to d were processed at the wafer level, with e and f at the die level.

Figure 3

Electromechanical performance of the fabricated G–AlN and reference devices. (a) Measured admittance curves and MBVD fitting of a G–AlN resonator and a reference device. The inset shows from left to right micrographs of a reference device with 100-nm thick Au top electrode and a graphene-electrode device. (b) Comparison between the measured operating frequencies of the fabricated 13 G–AlN NEMS resonators and the 15 reference devices and their FEM simulation results including an ideal metal-free resonator with the same W₀. (c) Q factor comparison.

Figure 4

Measured and simulated IR absorption spectra. Three solid lines are the spectra for different materials on top of 460 nm AlN and 100 nm Pt. The dashed line is the simulated spectrum for only the AlN–Pt stack. The inset shows the simulated electric field distribution of the fundamental mode of the Fabry–Perot resonance at 3.4 μm.

Figure 5

The thermal properties of the G–AlN resonator and reference device evaluated by 3D FEM simulations with applied IR power of 1 μW. The inset shows the simulated temperature distribution of the G–AlN resonator.

Figure 6

Experimental setup for the IR sensing measurement. Three reflective mirrors (M1, M2, and M3) and a dichroic filter are properly set up to co-align a red laser beam with a 5-μm QCL beam to facilitate the alignment between the QCL beam and the device under test.

Figure 7

Measured frequency response of the G–AlN and reference devices exposed to a 5-μm IR radiation modulated at 1 Hz by a chopper. The G–AlN detector showed a responsivity ~13× stronger than the reference device with 100 nm Au as the top electrode.

Figure 8

Absorption wavelength dependence on AlN thickness. (a) IR absorption spectra of various AlN thicknesses measured by FTIR. The insets from left to right show the simulated electric field distribution of the higher-order mode and the fundamental mode of the Fabry–Perot resonance, as well as a case of out-of-resonance for 510-nm thick AlN on 100 nm Pt, respectively. (b) Measured and simulated absorption wavelengths of various AlN thicknesses.