Unveiling the Design of MEMS Pressure Sensor Chips: From Thin Films to Resistors, Key to 0.1% F.S. Accuracy

The barometer in your phone and the tire pressure monitoring system in your car are both powered by a “compact yet precise” MEMS pressure sensor. It operates based on the piezoresistive effect of silicon, measuring only in millimeters, yet it can accurately sense pressure changes. This design document from Wang Dajun’s team breaks down the design of silicon pressure sensor chips based on MEMS technology: from the selection of the sensitive film shape to the calculation of the sensitive resistor size, and the precise positioning of the resistors, each step directly affects the sensor’s accuracy. Today, we will simplify the concepts to help you understand the “great intricacies” within this “small chip”.

1. Understanding the Basics: MEMS Pressure Sensors Sense Pressure through “Silicon Deformation”

The core principle of this sensor is quite simple —the piezoresistive effect: When single crystal silicon is compressed or stretched by external forces, its resistivity changes, and the resistance value made from it also changes. By connecting four such “force-sensitive resistors” into a “Wheatstone bridge”, the change in resistance caused by pressure will unbalance the bridge, outputting a voltage signal that can measure the pressure magnitude.

For example: When the film bends down under pressure, two resistors are stretched (resistance increases), while the other two are compressed (resistance decreases). The voltage difference output by the bridge is the “pressure signal”, with an accuracy of 0.1% to 0.25% F.S. (full-scale error), which means that for a pressure measurement of 100 kPa, the maximum error is only 0.25 kPa, smaller than the pressure of a single hair.

2. Core Design 1: Sensitive Film — Choosing Square over Round for Practical Reasons

The sensitive film is the “stress deformation area” of the sensor, akin to a “spring leaf”. Its shape and size directly affect sensitivity and linearity. During the design, comparisons were made between circular, rectangular, and square diaphragms, ultimately selectingsquare for three reasons:

  1. Manufacturing Ease: Square geometry is symmetrical, making it easier to cut according to the silicon crystal orientation, unlike circular shapes which require complex positioning;
  1. Sufficient Output: Under the same area and thickness, when the aspect ratio of the square diaphragm is controlled between 1:1 and 1:2, the theoretical output can reach its maximum, outperforming rectangular shapes;
  1. Balanced Performance: Although circular diaphragms exhibit the best maximum stress and deflection performance, the comprehensive performance (sensitivity, linearity, manufacturing difficulty) of squares is more suitable for mass production, making it the mainstream choice in the industry.

Specific dimensions are also crucial (referencing domestic manufacturing levels):

  • Overall chip: edge length 1000~2000μm, thickness 200~400μm (about the size of a grain of rice);
  • Sensitive film: edge length 500~1000μm, thickness 20~40μm (thin areas etched using MEMS technology);
  • If using SOI technology (silicon on insulator), the film thickness can be reduced to below 10μm, enhancing sensitivity.

3. Core Design 2: Sensitive Resistor Strips — Size and Doping Matter, or Accuracy Will Collapse

The four force-sensitive resistors are the “sensing core”. During design, three key issues must be addressed to ensure stable Wheatstone bridge output and minimal temperature drift:

1. Doping: P-type silicon doped onto N-type silicon must be precise

  • Doping Type: Choose “P-type silicon doped on the surface of N-type silicon” to form a PN junction, ensuring stable piezoresistive coefficients while isolating interference;
  • Doping Concentration: Control between 1×10¹⁸ and 1×10²⁰/cm³, with a junction depth of about 3μm, minimizing the temperature influence on the bridge’s zero output and sensitivity (small temperature drift). The square resistance (resistance per unit area) should be between 250 and 300Ω/□, suitable for subsequent size calculations.

2. Size: Width 10μm, Length 155μm, not arbitrary

The width and length of the resistors must be calculated considering “power consumption, heat dissipation, and manufacturing capabilities”. The core is to “avoid self-heating of the resistors affecting accuracy”:

  • Width (W): If the resistor is too narrow, it will heat up significantly; if too wide, it will occupy too much area. Based on “maximum power consumption per unit area of 5×10⁻³mW/μm²” and actual current of 1~3mA, the final choice is 10μm (commonly used in the industry is 5~20μm);
  • Length (L): Given the target resistance value of 5kΩ (excitation voltage 5V, current 1mA), square resistance of 270Ω/□, and width of 10μm, the length is calculated to be approximately 155μm, plus the “corner factor” (additional length at the resistor’s bend) to ensure accurate resistance values;
  • Key Requirement: The width and length of the four resistors must be consistent, and doping must be uniform; otherwise, the bridge’s zero point will drift, leading to measurement errors.

3. Structure: Open-loop Design for Easy Calibration

The bridge is not connected to feedback resistors, designed in an “open-loop form”. The benefits are twofold: first, it allows direct measurement of each resistor’s value, facilitating troubleshooting; second, it enables the addition of compensation circuits (such as temperature drift compensation) later to further improve accuracy.

4. Core Design 3: Resistor Position — Placing Them Correctly Maximizes Sensitivity

Resistors are not randomly placed on the film; they must be positioned in the “region of maximum strain” to ensure the most significant change in resistance value and maximum Wheatstone bridge output.

According to finite element analysis: When the square diaphragm is subjected to pressure, the maximum strain occurs at the middle of the diaphragm’s outer edge (for example, at the midpoints of the four edges of the square). The design follows two principles:

  1. Crystal Orientation Alignment: Resistors are arranged along the silicon <110> and <1ī0> crystal orientations, where the piezoresistive coefficients are maximized, allowing for the detection of both positive and negative strains (one stretched, one compressed), maximizing the bridge output;
  1. Distance Control: The distance from the center of the resistor to the edge of the diaphragm should ideally be 2.5% of the diaphragm’s edge length (for example, if the diaphragm edge length is 900μm, the distance would be 22.5μm), and should not exceed 9%, otherwise the strain will be too small, reducing sensitivity.

Final Design: The four resistors are arranged in a 3-fold layout (to save area), positioned within 550~604μm from the chip’s outer edge, with a gap of 5μm, ensuring maximum strain while avoiding short circuits caused by manufacturing errors.

5. Experimental Results: Chip Validation Achieves 0.1%~0.25% F.S. Accuracy

Following this scheme, the integrated circuit research institute produced chips, and the assembled sensor chips were tested:

  • Accuracy: 0.1%~0.25% F.S., meeting the high precision requirements of industrial and consumer electronics;
  • Process: Compatible with conventional MEMS processes (etching, ion implantation, diffusion), suitable for mass production;
  • Advantages: Small size (chip edge length 2000μm), fast response, minimal temperature drift, applicable in scenarios such as mobile phone barometers, car tire pressure monitoring, and medical devices.

Interactive Time

What devices have you encountered that use pressure sensors? Is it the tire pressure monitoring system in cars or the altitude measurement in smartwatches? If you were to design a MEMS pressure sensor, which performance would you most want to optimize — a smaller size or lower power consumption? Let’s discuss in the comments!

#MEMS Pressure Sensor #Silicon Piezoresistive Effect #Sensor Chip Design #Wheatstone Bridge #High Precision Sensing

Leave a Comment