1成果简介 

        高灵敏声学传感器件是人机交互和人工智能（AI）技术的核心组件。传统电容式麦克风和微机电系统（MEMS）麦克风虽然已在消费电子中广泛应用，但在远距离语音拾取、超宽带声学探测、低噪声生理信号监测等方面仍面临灵敏度瓶颈。二维（2D）材料因其原子级薄厚度和卓越的力学性能，理论上是最理想的声学传感振膜材料。然而，如何制备大面积、无缺陷、可自由悬浮的二维材料薄膜，一直是该领域面临的关键技术挑战。

        本文，北京大学刘畅教授、王恩哥院士、刘开辉教授等在《Nature Communications》期刊发表名为"Highly sensitive microphones based on large freestanding reduced graphene oxide membranes"的论文。该研究提出了一种压力辅助双重转移策略（pressure-assisted double transfer strategy），成功制备了直径达8厘米、厚度仅80nm的大尺寸自由悬浮还原氧化石墨烯（rGO）薄膜，其径厚比高达约10?。

        基于这一超薄膜，研究团队构建了一种超灵敏声学麦克风，实现了~500 μm/Pa的静态压力响应度和1 kHz下高达~115 dB的动态信噪比（SNR）。该rGO基麦克风具有100 Hz–50 kHz的超宽带频率响应，在9米远距离语音识别任务中，语言识别准确率达到90%，显著优于商用MEMS麦克风的70%。该工作为大面积自由悬浮超薄膜的可靠制备开辟了新路径，将有力推动先进声学器件的发展

        2图文导读  

        

        图1、 Fabrication process for a suspended rGO membrane with a diameter of 28 mm. First, a graphene oxide (GO) slurry was diluted with a methanol–water mixed solvent (volume ratio = 4:1) to prepare a GO dispersion with a concentration of 2 mg·L?1. The resultant dispersion was then spin-coated onto silicon wafers pretreated by air plasma. The coated wafers were immersed in concentrated iodic acid to reduce the GO membrane, yielding a rGO membrane anchored on the substrate. Subsequently, the rGO membrane was released onto the surface of an aqueous solution containing sodium dodecyl sulphate, which served to reduce the interfacial surface tension and facilitate membrane detachment. A polished larger 304 stainless-steel frame (with an inner diameter of 40 mm as an example) was used to carefully lift the suspended membrane from the water. After air drying, the rGO membrane adhered uniformly to the frame, forming a relaxed, non-stretched suspended membrane. Thereafter, a smaller stainless-steel frame (with a target inner diameter of 28 mm) was aligned underneath the larger frame at a defined spacing. Combined with a sealing frame, the assembly was used to fix the relaxed membrane and seal the gas chamber. The entire structure was mounted onto a custom-built chamber, which was pumping gradually at a controlled rate by a vacuum pump. Ultimately, driven by the differential pressure across the membrane, the relaxed membrane conformed tightly to the smaller frame, forming a successful freestanding rGO membrane with a precise diameter of 28 mm and a specific tension.

        

        图2、 Resonance frequency of the rGO membranes in air. a, Schematic of the experiment. b, Schematic in COMSOL. c, Normalized time domain signal obtained from the experiment. d, Normalized spectrum of the rGO membrane with the fundamental frequency of ~120 Hz determined by COMSOL and the experiment, respectively. e, Fundamental resonance frequencies of rGO membranes (1 nm-5 μm thickness, 28 mm diameter, 8 MPa stress) from the membrane model, COMSOL simulations, and the experiments, respectively. f, Fundamental resonance frequencies of different diameters under prestress of 8 MPa determined by COMSOL.  

        

        图3、Calculated velocity response of a circular rGO membrane as a function of thickness by using the proposed lumped element model. From top to bottom, the thickness of graphene diaphragms is 1000 nm, 100 nm, and 10 nm. It’s obvious that as the thickness decreases, the small peaks are suppressed, potentially because the higher-order modes are more readily damped by air coupling. 

        

        图4、The configuration of the dual-electrode microphone. a, Schematic of the device configuration. b, c, Photographs of the electrode (b) and the supporting frame (c).

        

        图5、Pressure?deflection curve of the membrane with the stress of 0.5 MPa.

        

        图6、Equivalent circuit of the freestanding rGO microphone with a transimpedance amplifier configuration.  

        3小结 

        综上所述，该工作通过创新的压力辅助双重转移策略，成功攻克了大面积二维材料自由悬浮薄膜的制备难题，制备出直径8厘米、厚度80 nm、径厚比~10?的大尺寸rGO自由悬浮薄膜。基于该超薄膜构建的声学麦克风实现了~500 μm/Pa的超高静态压力响应度和~115 dB的动态信噪比，并在100 Hz–50 kHz超宽带范围内保持优异的频率响应平坦度。在实际应用的远距离语音识别任务中（9米），rGO麦克风的识别准确率达90%，较商用MEMS麦克风（70%）提升了20个百分点，充分展现了其在人工智能语音交互、远距离通信和低噪声声学监测中的巨大潜力。该工作不仅为大面积超薄膜的制备提供了可靠的技术路线，也为下一代高性能声学传感器件的开发奠定了材料与器件基础。