|Packaging optoelectronic components for the telecommunications industry is getting more complicated as vertical – cavity surface – emitting lasers (VCSELs) move from R&D into mainstream use in fiberoptic communication. This is particularly true for original equipment manufacturers (OEMs) and contract electronics manufacturers (CEMs) that are either transitioning into the telecom arena from other markets or trying to keep pace with the telecommunication industry’s increasingly varied laser use.
When working with VCSELs, a “precising” station on the bonder’s pickup pedestal serves two purposes. It enables the VCSEL to be taken from the Gel Pak with an unheated tool without damaging the Gel Pak prior to handling with a preheated tool for placement on a submount. It also holds the VCSEL firmly in place so that when the pickup tool lands on the VCSEL its emission point is not moved off center from its alignment with the bonding system’s overlay image.
The bond head on the pickup tool must be either flat – faced with vacuum or a two – sided collet because the crystal facets on the opposite ends of edge – emitting diodes must not be touched by the bonder’s pickup tool. Touching can create flaws that cause faulty operation or even complete failure of the diode.
Since gold/tin alloy is the predominant bonding material used to attach diode lasers for the telecom industry, the pickup tool needs to be equipped with a dual head – one for the die and one for the gold/tin preform. Although not specified yet for telecom use, epoxy bonding of VCSELs, particularly in computer communications applications, is gaining favor with some packagers. For those applications, the bonding system design must include a special epoxy transfer tool for depositing precise amounts of epoxy.
Because of their tiny size (edge – emitters usually measure about 100µm thick by 200 to 300µm wide by 400 to 600µm long and VCSELs tend to measure about 300 to 400µm square), the bonder must be equipped with special viewing capabilities, such as this cube beamsplitter and scope, to allow the operator to first see the device and then, among other things, simultaneously view the diode laser image and the site in the submount where the die is to be placed.
To begin with, the operator cannot see the VCSEL’s emission point after the VCSEL has been picked up out of its carrier because the pickup tool holding the die blocks the view. While the laser’s emission point usually is placed on the center of the TO header, it frequently is located off center on the diode laser because of the need to make room on the laser chip for the bonding pad.
Therefore, during placement of the VCSEL on the submount, the emission point needs to be aligned to a specific feature on the submount. In some cases, the feature is a fiducial mark. In aligning the emission point to the center of the header, the issue is further complicated by the fact that the TO header is a machine – stamped piece of metal, for which tolerances may vary. For instance, if the TO header shoulder is out of round by a certain amount, it can adversely affect placement accuracy if not detected and corrected for.
In expensive and sophisticated automatic bonders, the pattern recognition system looks at the chip while it is sitting in the carrier and memorizes where the emission point is. But adding such capability to a semiautomatic bonder is not a viable option. Engineers at Semiconductor Equipment Corporation, therefore, use a relatively inexpensive but effective video image marker. It has enabled relative newcomers on the telecom scene such as Teledyne OptoElectronics (Marina Del Rey, CA) to achieve the desired +/-5µm placement precision of the emission point on any TO header. It also can be used for aligning to other types of submount surfaces.
The video image marker consists of a keyboard, a control box and X – Y knobs for adjusting the system’s magnified electronic images. The video image marker creates and stores a fixed video overlay of cross hairs in the form of a bit map for centering the VCSEL’s emission point, as well as up to 10 different video patterns that can be superimposed on the submount on which the VCSEL is to be bonded. An example of the latter would be a circle overlay image of the outside diameter of the top of a TO header (see Fig. 2). Custom patterns also can be created.
The bonder operator uses images/patterns from its video image marker to line up the VCSEL to the target bond site on the TO header. The video image marker is also used to enhance precision during die pickup out of the die carrier. It calibrates in metric units or inches and the measurements are displayed on the screen of the bonder’s color monitor.
|As appeared in Laser Focus World – May 2001|