Issue



X-Ray Analysis: Improving Inspection Quality


06/01/2005







NEW TECHNOLOGY QUICKLY IDENTIFIES FAULTS

BY VIKRAM BUTANI AND DONALD NAUGLER

The economic environment in the electronics industry has changed dramatically over the last two years. The focus on the production floor has moved from increasing capacity to improving the yield. One of the major consequences of this shift in focus is an increased interest in x-ray inspection.

Present Day Systems

Standard x-ray inspection systems depend heavily on the operator’s ability to manipulate the controls in order to obtain the best possible image. Further, once the image is obtained, the next hurdle is the operator’s ability to interpret the image and make decisions to improve the manufacturing process. Thus, the quality of the image obtained plays a critical role in the operator’s interpretation and eventually in the improvement of the yield.

Traditionally, x-ray inspection has involved manually adjusting contrast and brightness for different regions of the image. This is done to improve the visibility of low-contrast defects across different thicknesses and materials. Common defects that are inspected are voids in BGAs or voids in encapsulants for flip chip components. However, what the operator captures and analyzes is based on what the operator is able to see in the first place. Often, there are a number of extraneous factors, such as lack of experience, fatigue, work environment, etc. that will impact the operator’s ability on any given day. These result in a longer inspection duration and inconsistent decision-making from one operator to another, which have major cost implications in the production inspection context. Thus, an ideal scenario for maximum output from the x-ray tool is to obtain an image detailing all the features on a consistent basis, irrespective of the operator’s ability.

X-ray Detector Technology

The detectors in current x-ray systems fall into two main categories: analog and digital. Analog systems are either open or closed. Open systems consist of a vidicon camera with a phosphor plate attached to the front. A vidicon camera carries a tube whose charge density pattern is formed by photoconduction and stored on a photoconductor surface that is scanned by an electron beam, usually of low-velocity electrons (Figure 1). The reason an open system is called open is because the entire set-up is not enclosed in a vacuum tube. Open systems are economical, but suffer in resolution. These systems are useful for inspecting gross defects, but become inadequate with respect to micro-mil defects. A detector with a closed analog imaging system (for example, the image intensifier) is popular because, apart from giving better contrast in comparison to other analog systems, it is inexpensive and, due to its design, is easy to maintain. The largest component of the image intensifier is a vacuum tube, which explains why it is a closed system. This tube is convex on one end and is covered by a thin layer of cesium iodide (CsI). The cesium iodide forms the cathode of the detector, which first detects the x-rays. When the x-rays strike the CsI, they are turned into visible light, which is then converted to electrons. Once the electrons strike the P-20 (a light producing green phosphor) on the anode, they produce visible light. The output from the image intensifier enters an imaging chain consisting of lenses, a CCD camera, and the final output onto a monitor. The image intensifier output is real-time (which refers to the performance of a computation during the time of a related physical process, the live results of which help guide the physical process) at 30 frames/sec.


Figure 1. Example of a component using current x-ray detector technology.
Click here to enlarge image

The image intensifier, however, has a fair share of limiting factors. A distortion of 15 to 17% typically is visible because of the transfer of a flat image onto a curved surface. In comparison to the digital detectors available today, the image intensifier can be quite large and bulky. The image intensifier starts with an 8-bit image and, by the time the image goes through the imaging train and the A to D conversion, it is reduced to 4 to 5 bits, or about 32 grayscale levels. This limits the contrast capabilities of the image, thereby reducing its ability to capture subtle differences in the densities of the object (e.g., small voids or defects in plastic encapsulants). The image intensifier applies a phosphur film to convert x-ray energy into visible light, in the transfer there is a tendency for brightness caused by penetration of low-density materials to overflow onto the image of neighboring material of higher density. This leads to the occurrence of a phenomenon called “blooming,” or veiling glare. This phenomenon is more obvious in image intensifiers than in open systems not because open systems are better, but because their resolution is not high enough to pick up on the effect.

Amorphous silicon (a-Si) imaging technology, which was developed by medical equipment manufacturers for digital radiography, has been boosted by the recent breakthroughs in thin film transistor arrays similar to those found in today’s notebook computer screens. As a result, the latest generation of a-Si detectors, which can generate images in a digital 12-bit format, yielding more than 4,000 grayscales for analysis, have become efficient enough in achieving the resolution needed for a wide range of applications.

The a-Si technology is based on 2-D, solid-state, amorphous silicon imaging arrays that contain hydrogen. The arrays, which can be fabricated with an area up to 16 × 16 sq. in., provide about one million sensors. Combined with a CsI scintillator (phosphor screen), the sensor presents an ideal solution for high-resolution x-ray imaging applications. A scintillator is deposited directly onto the surface of the detector. X-ray photons striking the phosphor are converted to visible light, which is absorbed and converted to an electric charge by the photodiodes. The charge is integrated on each photodiode so that each pixel collects a signal proportional to the local flux of the x-ray beam. When the array circuitry scans the diodes, the charge is converted to a video signal, which reproduces the x-rays image. The signal is then read out in real time as a digital electronic image using thin film transistors made of the same amorphous silicon material. The image is then manipulated - it can be either read out and displayed continuously at 5 to 30 frames/sec., or integrated over many frames to be displayed at a frame every few seconds to improve sensitivity. In both cases the feedback to the operator is immediate.

New Imaging Technology

A new imaging technology* provides a sophisticated image enhancement tool that automatically optimizes x-ray images across density gradients for each region of an image (Figure 2). Since it performs the enhancement in a single pass, the enhanced image shows practically all defect structures, eliminating the need for manual x-ray output or detector, contrast, and brightness adjustments. This directly leads to substantial cost reduction, as the inspection period for each part is significantly decreased and consistency from one operator to another is dramatically increased.


Figure 2. Sample image of electronic component with and without automatic defect enhancement.
Click here to enlarge image

The new technology also results in higher inspection quality, because defect visibility is substantially improved. This aspect is particularly relevant as the material densities of substances being inspected get lower, as in the case of non-silver doped epoxy or optical fiber. These low-density materials, when assembled in components with other high-density material, are almost impossible to inspect with standard x-ray systems. This is due to the contrast depth of today’s analog detectors. The typical image intensifier has an 8-bit output, which goes through an A to D conversion leading to a 5-bit image depth or 32 shades of gray on the monitor. With such limited contrast resolution, it is difficult to capture subtle features at the lower fringes of the grayscale spectrum. These features, previously on the fringes of visibility, are now obvious to the engineer.

Conclusion

As customer requirements for x-ray capabilities increase, more sophisticated software tools have become available for the benefit of the production engineer. The new technology removes subjectivity from the inspection process and places less dependence on operator interpretation. This, in turn, ensures the engineer that the x-ray image is a true representation of the product being inspected. X-ray analysis can be accomplished in far less time, allowing the manufacturer to keep pace with ever-increasing productivity demands. This savings in time, plus the advent of unbiased image analysis, contribute substantially to improved yields and the evolution of leaner production environments in electronics manufacturing.

* VJ Electronix’s Automatic Defect Enhancement technology.

VIKRAM BUTANI, formerly with VJ Electronix Inc., and DONALD NAUGLER, general manager, may be contacted at VJ Electronix Inc., 1000 Mount Laurel Circle, Shirley, MA 01464; (978) 425-9446; e-mail: [email protected].