The semiconductor industry generally requires efficient and accurate detection of surface defects on wafers, which can capture effective defects and achieve real-time detection. The more common surface detection techniques can be mainly divided into two categories: needle contact method and non-contact method, with needle contact method as the representative of contact method; Non contact methods can be divided into atomic force method and optical method. In specific use, it can be divided into imaging and non imaging.
The needle touch method, as the name suggests, is a surface detection method that detects by the contact between the needle and the material being tested, and is one of the earliest methods in the manufacturing industry. The shape and contour information of the tested surface is transmitted to the sensor through the stylus, so the size and shape of the stylus are particularly important. According to the detection principle of needle touch method, it is only possible to detect the true contour of the measured object when the radius of the needle tip approaches 0. However, the thinner the tip of the stylus, the greater the pressure generated on the measured surface, and the stylus is prone to wear and tear, scratching the surface of the measured object. For coated surfaces and soft metals, contact detection can easily damage the surface of the tested sample and is generally not suitable for use.
In 1981, Binnig and Rohrer invented the Scanning Tunneling Microscope (STM). STM utilizes the quantum tunneling effect, with the needle tip and the surface of the object being measured as two poles. Use a very thin needle tip to approach the surface of the sample, and when the distance is very close, a tunnel junction is formed. The distance between the needle tip and the surface of the sample is kept constant, allowing the needle tip to move in three dimensions on the surface of the sample. The atomic height felt by the needle tip is transmitted to the computer, and after post-processing, the three-dimensional morphology of the surface of the tested object is obtained. Due to the limitations of STM, Binnig et al. developed an atomic force microscope (AFM) based on STM. AFM detects the attraction or repulsion between the needle tip and the specimen, making it suitable for both conductive and non-conductive materials.
Scanning near-field optical microscope (SNOM) utilizes the characteristics of the near-field near the surface of the sample being tested to detect its surface morphology. Its resolution can far exceed the limit of conventional microscope resolution (λ/2).
The commonly used imaging detection methods in the semiconductor industry currently include automatic optical detection, X-ray detection, electron beam detection, etc. Scanning electron microscope (SEM) is a microscopic object research tool invented in 1965. SEM uses an electron beam to scan a sample, causing secondary electron emission, which can produce an enlarged morphology image of the sample surface. This type of image is enlarged point by point and has a certain order. The advantage of SEM is its extremely high resolution.
The combination of X-ray non-destructive testing technology and digital image processing technology can perform high-resolution detection of internal connections in devices. Agilent has a high market share, with typical products including the 5DX system.
Automatic Optical Inspection (AOI) technology is a detection technique based on optical principles. It detects defects on the surface of samples through the motion of precision instrument platforms, image acquisition devices, and digital image processing technology. Its advantage is that the detection speed is fast. AOI equipment has developed rapidly in China in recent years and can be considered to have considerable market potential. AOI technology obtains images through CCD or CMOS sensors, converts them from analog to digital, transmits them to a computer, processes them digitally, and compares them with standard images
