Innovative optical imaging method seen to lower IC test costs

Dr. Caesar Saloma is a Professor of Physics at the National Institute of Physics, UP Diliman, Philippines.

What is the technology all about?

We are concerned with the accurate determination of defects in integrated circuits and the scale that we are talking about is in the scale of micron, or even smaller. One technique to do this is to use optical microscopy. Right now, the standard optical microscope that is widely used in the semiconductor industry does not have what we call "optical sectioning effect." In other words, it is difficult to determine the 3D distribution of a defect.

Also, in optical microscopy, if you rely on reflectance information, it is difficult to discern or differentiate between a metal and a semiconductor material because both of them have high reflectance values. The 1P OBIC method entails the scanning of a one photon light beam across the semiconductor device. The idea is, if you shine light on a semiconductor material, it will release current, and that current can be measured. In OBIC, you have light as a probe, and the output is an electrical signal. It is important to note that only semiconductor materials generate an OBIC signal. A metal does not.

Is OBIC widely used in the industry today?

OBIC is widely used today. The problem is that there is an inherent disadvantage with OBIC. The images produced have no vertical resolution, which means you cannot tell which vertical plane is higher or lower than another. Normally, in integrated circuits, the surface is not really flat. There are some grooves but the lack of vertical resolution prevents us from recognizing that these grooves exist.

How much would this cost if this is implemented by a company?

Because the requirement is a standard continuous wave laser, which could be a helium neon laser, an argon laser, or a semiconductor laser in principle, much of the cost will not be on the light source but only on the construction of the microscope. The cost will be less than $10,000 if you build one from scratch.

It is important to note that there is another technique which can, in principle, do 3D imaging of semiconductor defects. It is the 2P OBIC. The problem with 2P OBIC is that it is a nonlinear process and has a low probability event, so you have to use high peak pulse lasers in order to improve the probability of realizing the phenomenon. Therefore you normally use a femtosecond pulse laser, which is quite expensive. The light source alone costs more than $100,000.

We have already applied for a patent which is shared by the authors, the University of the Philippines and the Department of Science and Technology.

Do you think this technology will innovate the way defects are seen in ICs?

I think so. But we have to bear in mind that there are many types of microscopes, because defects are of different sizes. Some are using STM (scanning tunneling microscope), or near field microscope for defects in the scale of nanometers or smaller. Depending on the scale or the resolution that you want to see, if you are looking at the micron level, optical microscopy provides the best solution in terms of cost and speed of acquiring the image. Moreover, the hardware requirement of the technique is quite simple - you can easily adapt an existing optical microscope to make it compatible with our technique.

- By Anthony Capati and Marcelino J. Alejo Jr.

Electronic Engineering Times - Asia