INTRODUCTION

With the increased use of optical fibers for communications has come a need for a robust means of joining fibers to a substrate and device. Narrow core high-bandwidth fiber requires extremely tight fiber-device alignment tolerances to prevent drastic signal reduction. For applications in harsh environments, a small droplet of solder is often used to hold the fiber in alignment with a device. The fiber is positioned to maximize optical performance of the device, and then the solder preform next to the fiber is melted so that it flows across the pad and over the fiber. Solder joint geometry considered here is shown in figure 1.

  

Figure: Geometry of the solder joint used to align an optical fiber. $\theta _f$ and $\theta _s$ are solder contact angles on the fiber and substrate, and $\phi _p$ is the angle at which a deflected fiber passes through the solder droplet. (Not to scale.)

Unfortunately, during the melting and solidification of the solder, significant motion of the fiber is observed to occur, decreasing the transmission efficiency. Because the capillary force on the fiber scales as its diameter and its stiffness as its diameter cubed, a very thin fiber is likely to be significantly deflected by the capillary force. Furthermore, as the solder joint solidifies, its volume changes, and with it the shape of the liquid surface and the magnitude of the capillary force on the fiber. Finally, as the solidified solder makes contact with the fiber, the shrinkage during cooling pulls the fiber further out of alignment.

In order to solve this problem, after the fiber is positioned for maximum transmission, it is offset to compensate for estimated shift, with some improvement in fiber-device collinearity. Here it is hoped that mathematical modeling will bring about an understanding of the problem which will lead to more robust design and less shift, to enable use of higher-bandwidth fiber with tighter collinearity tolerances than are currently achievable.