A Laser Scanner for Semiconductor Devices by David E. Sawyer and David W. Berning Abstract: This is a construction guide and operators manual for a laser scanner built for semiconductor device studies. A very brief discussion of the theory of operation of the scanner is given. The scanner's operation from a systems point of view is described in detail with emphasis on block diagrams. The scanner is described from a hardware point of view with a detailed description of the function of the various controls on the electronic equipment that was built for the laser scanner. A quick guide for the use of the scanner is given so that a person unfamiliar with the instrument can use it effectively. Specifications relating to the scanner's data gathering ability are also given. Mechanical drawings and circuit schematics are given to enable others to build a similar scanner. The optics and their alignment are discussed. Various display modes including color are discussed to enhance operator viewing. Key Words: Electronic reliability, electronics; laser scanner; measurement method; mixer; optics; pre-amplifier; radio receiver; semiconductor device studies. 1. INTRODUCTION PURPOSE OF THE LASER SCANNER The National Bureau of Standards laser scanner is a versatile apparatus which can be used to study photo-responsive materials and devices on a point by point basis. The scanner is intended for specimens that have a physical size such that they can be viewed under a typical optical microscope and would require magnifications that are typically achieved by such microscopes. The scanning system, by its very nature, produces an output which is an electrical signal that can be displayed in real time on a cathode ray tube (CRT) or which might be processed electronically in some other manner. Specifically, the National Bureau of Standards laser scanner was assembled to determine the usefulness and versatility of such an approach for studying semiconductor devices such as transistors and integrated circuits, and the materials from which these devices were made. The laser scanner is an apparatus that is able to nondestructively map photosensitive semiconductor devices and materials by scanning them with low power lasers. In general, the specimen being scanned is its own detector. The scanning operation is intended to reveal information which can be used to gain insight as to how well a particular semiconductor device is working on a point by point basis. The scanner can be used to locate weak points and stress points in functioning devices, as well as failure locations on failed devices. The scanner might also be valuable for checking semiconductor materials for quality (1). The scanner, like the scanning electron microscope (SEM), requires interpretation of the displays generated. Unlike the SEM, the scanner is completely non-destructive to the device being studied. Several aspects of the usefulness of the scanner described in this report have been reported in the literature. These include mapping high and low frequency signal characteristics of transistors, spatially pin-pointing nonuniformities in thermal characteristics and linearity characteristics of transistors, revealing logic states in MOS and bipolar integrated circuits, and changing the logical state of elements in an IC (2,3,4,5,6). Of course, an important requirement for optical scanning is that the active areas in the device must be accessible by light and not totally covered by metal. The trend, however, in modern device design is toward interdigitated structures and polysilicon gates, so it is easier to obtain scanning results now than it has been previously. In certain types of devices, it is also possible to scan the device from the back side, thus eliminating any problem connected with top metal coverage. In interpreting the display, the total operation of the specimen must be considered. For example, the information gathered by scanning an integrated circuit is obtained by monitoring changes in power supply current to the device, and so all the effects of photogenerating a current in a particular transistor within the IC may need to be fully taken into account. In scanning the transistor, only a small increase in collector current and thus power supply current would be produced if the device were electrically isolated. However, if the scanned transistor serves as a signal source for others, then the supply bus signal will be a composite of the various device currents. In some cases, the effect of scanning a portion of a device is to reduce the bus quiescent current, and this produces a darkened region on the display While these considerations are important in explaining apparent photoresponse signal strength and polarity, only the elements that give a photoresponse will show in the display; so the situation is not as complicated as it may seem. Furthermore, specimens having a large number of transistors are usually digital and small changes in current passing through a given transistor influence few others. Even if the scanned IC were linear rather than digital, some degree of response isolation may be achieved automatically. For example, many linear circuits contain negative feedback loops, and these may compensate for perturbations introduced inside the loops. In hopes that semiconductor device manufacturers and users would find it desirable to have a laser scanner similar to the one developed at the Bureau, this publication has been prepared to serve as both a construction guide and an operator's manual. Highlights in this publication include mechanical and electronic schematics, examples of uses, and a guide to the operation of the scanner. 2. OPERATING MECHANISM AT THE DEVICE LEVEL A moving light spot, generated by a mechanically deflected laser beam, has the effect of generating current carriers in the semiconductor material. In the simplest case, this can cause a current in an unbiased junction if the light spot falls sufficiently close to the junction. This current can be measured in an external circuit connected to the device being scanned. As the light spot is swept across the device being scanned, the amplitude of this current will vary as the composition of the semiconductor device changes. This current variation is the information that is used to present a photoresponse map of the device being scanned. The current variations are simply amplified and fed to a cathode ray display which is driven in synchronism with the laser scan. The information is obtained by interpreting the map produced. The system works equally well for mapping reversed-biased junctions. In this case the light spot, in generating carriers, causes the reverse current to increase. This change of current can be produced in the same manner as for the non-biased junction. If a transistor biased in the active mode is observed using the laser scanner, a change in basecollector junction current may be amplified by the transistor being scanned to give an enhanced photoresponse map. Another category of studies that can lend itself to observation on the scanner is that of resistivity. In particular, a resistor in an integrated circuit can show a photoresponse when the circuit is scanned. Again, the light spot creates carriers in the semiconductor material, but in this case the semiconductor becomes more conductive and the current passing through the resistor increases. For the resistor to be detected on a photoresponse map of the device being scanned, the resistor must be carrying a current. The scanner can be used to tell how uniform a resistive area is. If the resistor is in the base circuit of a transistor biased in its active region, this current can be amplified to give an enhanced photoresponse. In FET transistors and integrated circuits the light spot can cause an increase in the current in the channel. With proper interpretation one can tell which transistors in MOS arrays are on and which are off. 3. PRINCIPLES OF OPERATION 3.1 Optical and Mechanical Operation The laser scanner incorporates two low-power, continuous wave helium-neon lasers. One of the lasers operates at a wavelength of 0.633 μm in the visible portion of the spectrum, and the other at a wavelength of 1.15 um in the near infrared. A movable slide in the optical paths of the lasers allows the user to select which of the lasers is to be used for a particular experiment. The chosen laser is sequentially reflected from two nodding scanning mirrors which oscillate in orthogonal directions. The moving light spot forms a raster on the specimen upon being directed through the camera tube of an optical microscope (see Appendix A for a description of the microscope). specimen is electrically connected to a cathode ray display which is deflected in synchronism with the scanning mirrors. The The Let us examine the system in more detail. Figure 1 is a pictorial diagram of the light and signal paths of the dual-laser scanner. beam from the 1.15 um laser is deflected from M, through a half wave plate onto M2. The half wave plate is used to rotate the polarization angle of the 1.15 μm radiation to maximize the intensity of the radiation at the specimen and, more importantly, at the reflected light photodiode (see Appendix A). The mirrors M1 and M2 fold the optical path to make the scanner more compact and rugged. Mirror M2 directs the 1.15 μm laser beam toward the first deflection mirror labeled V. The beam of the other laser, which operates at 0.633 μm, passes through a polarizer-analyzer, and is deflected from mirror M4 onto mirror M3. The analyzer can be used to attenuate the beam when desired. Mirrors M4 and M3 are located on a movable slide. When this slide is moved to the right, the beam is reflected from M3 onto the first deflection mirror V. When the slide is in this position the beam from the 1.15 um laser is blocked. When the slide is moved to the left, the 1.15 um radiation is allowed to fail on deflection mirror V and mirrors M4 and M3 no longer direct the 0.633 beam onto V. The first deflection mirror V generates the vertical component of the scanning raster. This mirror is a commercially available galvanometer device. This mirror is driven with a triangle waveform through a suitable power amplifier. After the light leaves the vertical deflection mirror V it passes through a lens system designated L1 which has the effect of focusing the vertical deflection from V onto the |