1. Introduction Atmospheric pressure sensors play an irreplaceable role in industrial production, weather forecasting, climate analysis, environmental monitoring, and aerospace. Traditional pressure sensors are generally mechanical, resulting in large sizes that hinder miniaturization and integration. MEMS technology not only overcomes these shortcomings but also significantly reduces costs while offering superior performance. Currently, widely used MEMS-based pressure sensors fall into two main categories: piezoresistive and capacitive. Piezoresistive pressure sensors exhibit excellent linearity but generally low accuracy, large temperature drift, and poor consistency. In contrast, capacitive pressure sensors offer higher accuracy, smaller temperature drift, and a more robust chip structure, but suffer from poor linearity and are susceptible to parasitic capacitance. Currently, MEMS capacitive pressure sensors are primarily used for overpressure measurement, with limited applications in meteorological pressure measurement due to their high cost. Therefore, this paper presents a high-performance, low-cost miniature capacitive meteorological pressure sensor. The entire process is simple and standardized, using single-crystal silicon as the thin film material. A contact structure is employed, utilizing anodic bonding to form a vacuum cavity, followed by anisotropic etching and deep etching with KOH to form the silicon thin film. Experimental results demonstrate that this sensor is suitable for meteorological pressure measurement. 2. Basic Principles and Structure The basic structure of a capacitive pressure sensor is shown in Figure 1. Where: ε0 is the dielectric constant in vacuum; t is the thickness of the insulating layer; εr is the relative dielectric constant of the insulating layer; g is the initial distance between the two plates of the capacitor under zero load; ω(x, y) is the vertical displacement of the mid-plane of the plate film. As the formula shows, external pressure changes the capacitance by altering the area and spacing of the capacitor plates. As the pressure gradually increases, the capacitance increases due to the decrease in the plate spacing; at this point, the capacitance value is determined by the non-contact capacitance. When the two plates are in contact, the capacitance is mainly determined by the contact capacitance. 3. Sensor Design and Manufacturing The sensitive film is the core component of the sensor; its material, size, and thickness determine the sensor's performance. Currently, the materials used for sensitive films are mostly heavily doped p-type silicon, Si3N4, and single-crystal silicon. Each of these materials has its advantages and disadvantages, and their selection depends on the target requirements and specific processes. Silicon films do not disrupt the crystal lattice, have excellent mechanical properties, and are suitable for anodic bonding to form cavities. From the perspective of simplifying the process, this scheme selects silicon films. The working state of the contact-type thin film was simulated using the finite element analysis software ANSYS. The material was Si, the film was square with a side length of 1000 μm, a thickness of 5 μm, and an electrode spacing of 10 μm. Under atmospheric pressure of 1.01 × 10⁵ Pa, the central contact portion and the four corners of the film were essentially stress-free, with the maximum stress at the center of the four sides being 1.07 MPa, less than the yield stress of silicon (7 MPa). The stress distribution is shown in Figure 2. The entire manufacturing process followed standard procedures, as shown in Figure 3. First, 100 nm of SiO₂ was thermally oxidized, serving both as a mask for etching Si and as an insulating layer for the two electrodes of the capacitor. Anisotropic etching was used to form the capacitor cavity and the stop grooves for exposing the electrodes. If the silicon wafer thickness was uniform and the KOH etching rate was uniform, this method could be largely equivalent to self-stopping etching. The two electrodes of the capacitor were led out from the glass and then anoly bonded to the silicon wafer. After thinning the bonded sheet using KOH etching, the reactive ions deeply etched to expose the measuring electrodes. 4 Key Processes 4.1 KOH Anisotropic Etching Among various anisotropic etching methods, KOH etching is simple, practical, and inexpensive. However, in cases of large-area, deep etching of silicon wafers, KOH etching can easily affect the shape and smoothness of the wafer surface. Therefore, selecting the appropriate solution ratio plays a crucial role. Under conditions of KOH mass fraction of 20%–40%, silicon wafer resistivity of 0.05 Ω·cm, and constant temperature in an 80℃ water bath, the corrosion surface undergoes significant changes with increasing KOH concentration: the raised mounds gradually change from cones to octagonal pyramids and then to square pyramids, as shown in Figure 4(a). The height of the pyramids is mostly tens of micrometers, and the base length is one or two hundred micrometers. With increasing KOH concentration, the mounds disappear, and square frustums appear, as shown in Figure 4(b). The depth of the frustums is mostly several micrometers, and the base length is one or two hundred micrometers. With further increasing KOH concentration, the shape of the pits changes. The complete square frustum pits almost disappear, and mostly sloping semi-square frustum pits appear, as shown in Figure 4(c). The slope height is 1–2 μm, and the side length is within 10 μm. The four sloping faces of the square pyramids and square frustums correspond to the (111) series crystal planes with the lowest corrosion rate. When the concentration is low, the corrosion rate ratio of the (100) and (111) crystal planes is small, so small hills appear; when the concentration increases, the corrosion rate ratio of the (100) and (111) crystal planes increases, so small pits appear; after the concentration reaches a certain level, the corrosion rate ratio of the (100) and (111) crystal planes tends to stabilize, and pits still appear, while the corrosion rate ratio of the (110) and (111) crystal planes increases, thus producing slopes. Only by adjusting the concentration of KOH to obtain matching corrosion rates of the (100), (110), and (111) crystal planes can a better corrosion surface be obtained. The experiment also shows that temperature mainly affects the corrosion rate and has little effect on the corrosion morphology of silicon wafers. 4.2 Anodic Bonding Currently, vacuum cavities are mostly formed using Si-Si bonding or anodic bonding. This scheme uses anodic bonding because anodic bonding has lower requirements than Si-Si bonding. First, the temperature only needs to be 400-500℃, and second, the surface finish requirement is also relatively low. This process involves metal electrodes, making high temperatures unsuitable. The bonding surface has electrode leads approximately 1400 nm high and 20 μm wide. After a certain degree of anisotropic etching with KOH, the roughness of the SiO2 on the bonding surface is about 100 nm. Experiments have shown good bonding (Figure 5) and a good sealing effect. 4.3 Reactive Ion Deep Etching Reactive ion deep etching (DRIE) can create very deep vertical structures. This experiment used it for the final silicon thin film formation. The etching effect of DRIE (depth of 250 μm) is not as flat as KOH etching; the etched surface is relatively rough, with surface particle undulations of several micrometers, as shown in Figure 6. Furthermore, etching non-uniformity exists; the electrodes are etched to the exposed edges of the 75 mm silicon wafer, while the electrodes in the center of the wafer are not yet exposed. The non-uniformity of deep etching is closely related to the pattern of the etched surface, but the causes and mechanisms are currently not explained by specific and reasonable theories. Therefore, it is impossible to theoretically guide the design of the etched surface shape; it relies more on manual adjustment based on experience. 5. Experimental Results and Analysis The fabricated sensor sample. The thin film size is 2 mm × 2 mm, and the theoretical film thickness is designed to be 10 μm. However, due to the inherent ±20 μm thickness fluctuation error of the silicon wafer itself, and the difficulty in ensuring the design requirements after KOH anisotropic etching and reactive ion etching, the actual film thickness varies from 10 to 30 μm. The pressure sensor was measured at room temperature (19.34℃). The measuring device was a Druck DPI610IS, and the measuring circuit used an AD7745 capacitance measurement chip from Analog Devices, achieving an accuracy of 4 fF. The measurement curve is shown in Figure 7, with a test accuracy of 8.1‰. Due to the relatively thick silicon thin film, the linear portion within the measurement range is limited. Furthermore, the low area utilization of the capacitor electrode results in a small change in capacitance. These are the main reasons for the low performance. However, the figure shows that the measurement curve exhibits good consistency and repeatability. 6. Conclusion Utilizing the excellent mechanical properties of the silicon film, a capacitive pressure sensor sample was fabricated using a contact structure and a simple, standard process. Testing and analysis of the sensor have proven its applicability in meteorological pressure measurement. Further research will focus on improving its structural design and manufacturing process to enhance measurement accuracy.