Optoelectronic Packages Demand Advanced Inspection

Those of us even remotely familiar with the overwhelming potential of the optoelectronics industry are sometimes frustrated with its seemingly sluggish pace. Now boasting device speeds of up to 40 Gb/s, we find ourselves wondering how it is that light has not completely eradicated archaic copper already. One answer is to be found in a facet of IC fabrication that is commonly overlooked and underestimated, device-packaging complexity.

The goals of packaging for most non-optical IC devices are relatively straightforward. First, there must be a "home" for the device that is mechanically stable and thermally friendly with suitable electrical access. Second, the device must be protected from electromagnetic interference.

Optoelectronic packaging goals include these aims plus more stringent criteria. There must be provisions for optical access — getting light into and out of the package. In addition, much tighter tolerances and reliability requirements are usually necessary than those associated with standard IC packages. Optoelectronics packages are often hermetically sealed to protect waveguides and lenses from condensation. Electrical and thermal stability are more crucial when dealing with the extremely high-speed data transfer rates, while noise and impedance are particularly poorly tolerated in these applications.

Moreover, optoelectronic devices themselves continue to increase in complexity. The simplest model of such a device, a laser diode or light emitting diode (LED) beaming light to a photodiode through a fiber optic, is a far cry from the complex mechanisms currently on the market. As we build ever more complex optical networks, simple transmit and receive functions are no longer sufficient. Modules must be incorporated into these devices to amplify, modulate, switch, split, variably attenuate and couple the signals. Virtually any imaginable function leads to a new optoelectronic component. A good "boilerplate" model for envisioning one of these devices is shown in Figure 1.

1. Cutaway view of typical optoelectronic package

As a host of different devices exist, individual modules assume different forms to suit given applications. If the device contains a light-emitting source, the active element may be as simple as a LED, or more commonly, one of a variety of lasers (Table 1). Devices utilizing a light-detection element may utilize a photodiode, semiconductor optical amplifier (SOA), various waveguides and some rather extravagant optical microelectromechanical (MEMS) devices.

The "skin," so to speak, of these packages consists of a metal, ceramic or polymer casing. This choice is largely dependent on the device's thermal, mechanical and optical requirements. For heremetic access, there may be simply a clear, polymeric or glass window into the device. Typically, the more complex devices must sustain a host of supporting electronics. Then, thermal stability becomes an important consideration. To counteract thermal effects, custom-designed thermoelectric coolers are integrated onboard to maintain a stable operating temperature.

Invariably, these devices serve a number of different functions. A good sampling of the most common photonic devices currently packaged is also depicted in Table 1.

Efficient optical coupling encompasses many of the design challenges in designing an optoelectronic package. In these applications, coupling comes in three varieties: coupling light from a source to a waveguide (that is, sending a beam of laser light through a fiber optic); coupling light from the waveguide to the detector; and coupling light from waveguide to waveguide. The first case is usually remedied through lensing with 1 or 2 µm tolerances. The second case poses an easier problem, as the light is thrown into a "bucket" where no lensing and less stringent alignment is necessary. However, waveguide-to-waveguide coupling presents some of the more difficult challenges in package design. Such a problem usually mandates the use of ball lenses, lensed fibers or graded index lenses. Hermetic access becomes necessary as well, requiring the placement of windows, larger lenses or metallized fiber optics in the device.

Packaging costs, by far, dictate the highest fraction of the total cost of an assembled optoelectronic device. Components such as laser diodes or modulators, designed for high performance applications, are single-mode devices. They must be connected together using optical fibers or other types of waveguides with sub-micron alignment accuracies. Presently, this is accomplished mostly through "active" alignment, whereby a highly skilled technician performs sub-micron alignments through a microscope. Connections are made using epoxy, solder, laser welding or other attachments. Great care is taken that the alignments are not disturbed by the bonding process. This labor-intensive approach results in an output of only a few units per day in some cases, many times with non-uniform device characteristics. "Passive" alignment is sometimes used; the operator relies on sufficiently accurate part geometries for optical alignment. For maximized efficiency, most devices utilize a combination of the two strategies.

As with any product, reliability requirements dictate many of the design considerations involved in optoelectronic packaging. Reliability is bound to vary over market segments and customers. In this particular case, submarine cable is the most stringent, followed by terrestrial telecom and lastly, datacom applications.

X-ray technology works on the concept of penetration of the samples under inspection. The higher the material density of the sample, the higher the kV level required. However, at high kV levels, images of low-density materials will get washed out. Hence, the material content of optoelectronic packages makes optoelectronics a challenge.

Optoelectronic components consist of optical fibers with low material density combined with the III-V family of materials. III-V materials are binary crystals with one element from the metallic group 3 of the periodic table, and one from the non-metallic group 5. The family includes gallium arsenide (GaAs), indium phosphide (InP), gallium nitride (GaN), indium antimonide (InSb) and indium arsenide (InAs).

Traditional X-ray inspection uses image intensifier detector systems. These systems have an 8-bit analog output. This output goes through 'A to D' conversion and loses some signal to inherent noise. The image at the monitor is close to four to five bits of data or 16-32 shades of gray. Consequently these systems lack the range to capture and reproduce high-contrast resolution images of samples with a large variation in material density. Irrespective of the focal spot size on the source, subtle changes in material density are lost to image bloom or veiling glare. Thus, the brightness from the low-density regions bleeds over into the high-density areas. Or, to put it in another way, the high kV levels used for inspection of high-density material causes the images of low-density regions to be washed out.

Recently, the X-ray industry has started using a digital detector system for advanced inspection of high contrast applications. Amorphous silicon (a-Si) imaging technology, developed by medical equipment manufacturers for digital radiography, has gained importance by recent breakthroughs in thin film transistor arrays, similar to those found in notebook computer screens. As a result, the a-Si detectors, the latest generation that creates images in a digital 16-bit format yielding over 65,000 shades of gray for analysis, achieve the resolution needed for optoelectronic applications.

The technology is based on a two-dimensional, solid-state, amorphous (non-crystalline) silicon 'imaging array' that contains hydrogen. The arrays, which can be fabricated up to an area of 12 × 16 square inches, contain about one million sensors. Combined with a cesium iodide (CsI) or Lanex(r) 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 arrays. 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 into a video signal, which reproduces the X-ray image. The signal is read out in real time as a digital electronic image using thin film transistors made of the same amorphous silicon material (Figure 2).

2. Advanced X-ray inspection system

The image is then manipulated. It can either be read out and displayed continuously at 5 to 30 frames per second, 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.

In order to corroborate this hypothesis with some data, an experiment was conducted. Two different detectors were set up across from a microfocus X-ray source and a five-axis manipulator, one in an image intensifier system and the other in an amorphous silicon detector X-ray inspection system (VJ-2000 DIG system).

The sample inspected was a Kovar tube soldered with Sn62 alloy, first without and then with an optical fiber running through it. As can be seen from several images, the amorphous silicon detector system captured the large variation in density between the solder and the optical fiber in the same image. However, the image intensifier system produced a washed out image because it could not capture the low-density fiber while imaging the higher density solder material.

Though amorphous silicon systems have been used for several years by Universal Instruments, Indium Corporation, SMTC, Ford Motor Company, Boeing and others, few X-ray vendors in the electronics industry have extensive experience with these detectors. However, the growth of optoelectronics has created new demands on suppliers of X-ray inspection systems. Previous methodology using image intensifier systems is inadequate to the range of material densities that exist in optoelectronics applications. Now, X-ray systems based on amorphous silicon technology are able to create clear images of these complex components.