Configuring an X-ray Inspection System
By Udo E. Frank
Semiconductor components, high-density circuitry, wafer-level chip scale packages, stacked assemblies, micro-electromechanical systems (MEMS), and micro-opto-electromechanical systems (MOEMS) – the advantages of X-ray inspection are well recognized and understood for such applications. However, selecting the right X-ray system for each application can be a challenge.
A variety of factors come into play, including how equipment will be used and where it will be installed (in a production line or off-line), feature-size capability required, and a host of performance characteristics, such as contrast, sharpness, magnification, and acceptable noise level. Applications also can change. Selecting an X-ray inspection system often means purchasing equipment that exceeds current requirements.
An alternative, cost-effective approach is to purchase a basic system that is configured to meet current performance requirements. What distinguishes the system is the versatility of the platform concept, which allows the system’s X-ray inspection capabilities to be tailored to specific inspection needs. In this case, the basic configuration includes:
- Open microfocus tube with a transmission target – the tube features an acceleration range up to 160 kV and a detail detectability of 1 µm. The tube also incorporates a design that enables controlled and continuous stable output intensity for X-ray emission, constant image contrast, and brightness.
- Geometric magnification up to 2000×.
- 4-axes manipulator capability of accommodating sample sizes up to 440 × 550 mm (17 × 21 in.).
- 4-in. (102-mm) dual-field image intensifier.
- Digital-imaging chain with a CCD camera.
- A 17" (432-mm) monitor, GUI (graphical user interface) and real-time image processing system.
This system provides a basic X-ray imaging capability contained within a radiation-secure cabinet with a 1 × 1-m footprint. Additional capabilities can be added to the platform. Instead of a microfocus tube, a “multifocus” tube can be substituted. This tube offers three mode capabilities with a single tube: microfocus, nanofocus and high-power. Instead of a four-axes manipulator, a six-axes subassembly can be incorporated. Should the application warrant additional costs, add a high-resolution scientific-grade camera, BGA module, voiding calculation software and direct digital detectors.
With such a system, the decision comes down to what level of performance is required. What detail detectability, for instance, is necessary? If above 1 µm, then a microfocus tube will suffice. On the other hand, for improved feature recognition, a tube with nanofocus capability is required to meet these requirements. The decision will impact cost, because a microfocus tube costs less than a multifocus tube with an additional nanofocus mode. However, a cost/performance analysis based upon immediate and near-term application needs could determine that the extra, upfront expense of a multifocus tube.
Design and Performance Considerations
X-ray systems primarily consist of three subassemblies: an X-ray source, remote- controlled fixture for holding and manipulating the sample, and a radiation detector (Figure 1). These sub-assemblies are contained in a multiple-fused, radiation-shielded cabinet.
 Figure 1. Diagram of a typical X-ray system for industrial inspection. |
The X-ray source for industrial inspection is a tube, which may be “sealed” or “open.” An open design usually is a stainless-steel tube, in which a continuous vacuum is created, and it can be opened for cleaning and maintenance. However, a sealed tube is one in which the vacuum is introduced at the time of manufacture, and it cannot be opened without destroying it.
The manipulator is designed to provide precise X-Y-Z positioning and rotating/tilting of the sample at varied speeds, depending on whether a quick overview of the part at low magnification, or a more detailed examination at higher magnification, is desired. The function of the detector is to convert the real-time X-ray data into an image of visible light that can be seen and examined by the human eye. The most common detector is a combination video camera/image intensifier. Other types include high-dynamic scientific-grade cameras and flat-panel direct digital detectors (DDDs).
To determine the required configuration for a particular application, choices are required in terms of the subassemblies to be installed in the cabinet, such as type of mode required: microfocus, nanofocus, or high-power; most appropriate type of tube: open or sealed; what motion is required of the manipulator; and what type of imaging system is most suitable. These are a few of the specific design questions that should be considered:
 Figure 2. Microfocus transmission tube showing generation of X-ray beam. |
Contrast. X-rays are generated within a continuously evacuated tube housing. Electrons are emitted from a cathode, accelerated through a high-tension field (typically 10 to 225 kV) and focused onto a thin layer of target material, usually tungsten (Figure 2). As electrons collide with target-material particles, they are slowed down, and the loss in kinetic energy is converted into other types of energy, mostly heat, but also X-rays. Acceleration voltage determines the velocity and violence of the collisions with the target material; and therefore, the penetration power of the X-rays generated. At 30 kV, more soft radiation, which is highly absorbed in typical electronics assemblies, is generated. At 160 kV, more hard radiation is generated. The objective in selecting acceleration voltage for a particular application is to achieve the highest possible absorption (or attenuation) difference. Thus, in viewing a solder ball with a void, radiation must penetrate the ball sufficiently to depict the difference in absorption between the solid metal and the less-dense area of the void. Therefore, finding the optimal contrast is a one-parameter task of determining proper acceleration voltage, because all other parameters in an X-ray system can be adjusted automatically.
Sharpness. The focal spot, a resolution tube generally half the diameter of the focal spot, determines image or geometric sharpness. Therefore, a microfocus X-ray tube provides a resolution of about 1 µm. Only open tubes can achieve such high-resolution grades, are capable of higher-acceleration voltages than sealed tubes, and provide higher tube power with higher intensities (dose rates). They are the standard in X-ray microscopy.
 Figure 3. Relationship of focal spot diameter and geometric sharpness. |
As seen in Figure 3, the near-punctiform shape of the focal spot eliminates peripheral shadowing in the X-ray image almost entirely, achieving a nearly infinite in-depth sharpness. The required resolution is important when selecting tube types for the application. The smaller the focal-spot capability, the higher the tube cost, and vice versa.
Magnification. All absorbing structures of the 3-D sample under inspection are projected as a 2-D shadow image onto the entrance field of the detector. Geometric magnification depends on the position of the sample between the X-ray source and detector plane on the relationship of the focus-to-detector distance (FDD) and focus-to-object distance (FOD). Magnification (m) equals FDD divided by FOD. Magnification increases significantly within the last few millimeters toward the focal spot of the X-ray source. Because of this, the X-ray tube must be a transmission-target type, as a transmission target features a thin X-ray window (250 µm) located practically at the same level with the focal spot and a thin target layer sputtered onto the inner side of the X-ray window from which X-rays emit. With a typical FDD of 500 mm, maximum geometric magnification of 2000× can be achieved if the sample to be examined touches the outer surface of the X-ray window. This also is called “stamp magnification.”
An FOD of 1.25 mm results in a geometric magnification of 400×, while an FOD of 5 mm achieves 100×. For a sample that is 5-mm high, total achievable geometric magnification can vary by a factor of 20. In selecting a tube design for a particular application, maximum geometric magnification required must be taken into consideration.
Total magnification is visible to the operator on the displaying media. If, for example, the X-ray image is copied from a monitor screen onto a wall (by an LCD projector), the image on the wall is enlarged visibly compared to that on the screen. However, the resolution remains the same because only the image pixels are displayed larger in size.
Noise. In addition to image contrast and sharpness, noise is an important factor in determining image quality because it affects the observed clarity. For optical inspection, noise plays an inferior role because sufficient photons are available. For X-ray inspection, however, the situation is different; X-ray intensity, or dose rate, has a significant influence on noise. The relation of absorption signal to the gray-value noise level (frequently referred to as the signal-to-noise ratio) doubles with quadrupled intensity. The objective in X-ray inspection is to achieve the highest possible intensity. An Isowatt function keeps electrical power applied to the target constant and ensures consistent optimal conditions independent of the selected acceleration voltage.
A recursion filter also is an efficient tool for reducing noise level without losing real-time impressions of an image. It creates an output image that is the weighted sum of prior images in a time sequence. During imaging, gray values of individual captured images are being added pixel-wise, and averaged. Images that were captured earlier will be less weighted than those captured more recently. Noise reduction can be achieved for an unmoved sample. Should the sample be moved; however, the image would be “smeared.” This can be avoided when reducing the recursion level during movement so the image appears to be noisier, while the sample structures remain clearly discernible during motion. An automatic adjustment of the recursion level to the motion status (“moving noise reduction”) ensures user-friendly operation.
FDD also influences X-ray intensity. The relation is an “inverse square.” Therefore, when the FDD is doubled, the captured intensity is reduced to a quarter of the previous amount. For this reason, the sample should be positioned as close to the X-ray tube as possible, while adjusting the FDD to the required magnification grade. This is possible only with systems enabling a variable-detector position.
Leaving the pixel number constant where the converted entrance image is imaged results in useful post-magnification effects. This is possible when the detector unit consists of a combination of an image intensifier and camera. The image intensifier converts X-ray waves into visible light and amplifies them. The optical image available at the output field is observed by a camera and displayed onto a computer monitor. The most cost-efficient solution is a video camera (760 × 570 pixels with 256 gray values = 8 bits), which is sufficient for a range of applications. The entrance field of the image intensifier can be set to various sizes electronically. Because the image size in the output field of the image intensifier and the pixel number of the camera image remain constant, magnification grade of the image displayed on the implemented system monitor changes. This feature is called post-magnification or image-intensifier zoom.
When switching the entrance image of the image intensifier from 4-in. diameter to 2 in., post-magnification grade is doubled while signal-to-noise ratio is decreased to half (“double noise”), because the entrance field of the detector is reduced to a quarter of its size. Using post-magnification effects is particularly useful when geometric magnification cannot be increased further, because the sample already touches the X-ray tube.
For higher requirements, a camera with more than 1-M image pixels is recommended in combination with an image-processing chain featuring a gray-value resolution of 16 bit = 65,536 gray values. This Scientific Grade Camera (SGC), which also is used with an image intensifier, offers the opportunity to reduce the image repeat frequency of 25 frames/second (real-time) to less than 1 frame/second. The advantage is in the acquisition time. The longer an image is exposed, the longer the X-rays can work in, the more X-ray photons can be captured for the image, and the lower the noise level.
For high requirements, a DDD is recommended with the highest pixel amount (1,888 × 1,408 pixels), best contrast resolution (16-bit), and lowest noise level because it enables real-time operation with 30 frames/second, while providing an image quality that can compete with an X-ray film. Moreover, in comparison to an image intensifier/camera combination, the DDD is about ⅛ the height, so the FDD and the maximally achievable magnification of the X-ray system can be increased considerably. The entrance plane of the detector is 24 × 18 cm, so that Eurocard Format (16 × 10 cm) PCBs can be displayed completely at 1.5-fold magnification. Because the pixel size of this detector is 127 µm, 85-µm structures can be resolved.
Conclusion
Inspection and analysis of electronic assemblies and parts using high-resolution X-ray technology mandates several choices when selecting a system. The tendency is to purchase a system with capabilities that exceed the need because of the minimum level of capability available, or in anticipation of future requirements. The result is a less-than-cost-effective purchase.
Another choice is available – the ability to configure a versatile X-ray platform that meets specific inspection needs. The versatile X-ray platform offers a cost-effective approach to purchasing the most suitable system. Tthe decision that must be made is one of determining the basic level of performance required. The type of tube selected represents the most fundamental decision, as it determines resolution and focal-spot size. Contrast, sharpness, magnification, and noise, as well as imaging requirements for the application, also represent an important set of choices. The advantage of such a system is significant – the likelihood of a more profitable bottom line.
UDO E. FRANK, Ph.D. has served as director of technology development at FEINFOCUS. For more information, please contact COMET North America Inc., 203/969-2161; e-mail: [email protected].