Optical Fiber Alignment

Precise fiber alignment is necessary for accurate and reliable data transmission in an optical network. Most optical networks have many optical couplings and even minor (< 1%) losses at these couplings accumulate to produce significant signal loss and consequent problems in data transmission. Minimizing coupling losses is critical in these networks. Good fiber alignment produces the highest coupling efficiency and therefore the least signal loss before assembly or packaging of an optical system. Minimal signal loss results in reduced power requirements which, in turn, means fewer repeaters, lower investment costs, and reduced incidents of failure.
Coupling between laser diode and fiber
Figure 1. Coupling between laser diode and fiber.
The basic concepts of fiber alignment are illustrated in Figure 1. A well-characterized input beam (in this case from a laser diode) is coupled into the fiber under test and a raster scan of the fiber is used to detect first light - i.e., the output signal from the fiber indicating when the laser beam first enters the fiber. Once first light is detected, the position of the fiber is adjusted in a lateral, longitudinal, and angular coordinate system to locate the peak intensity of the output optical signal. In the simplest case, only lateral (X, Y) adjustments are necessary, while in multi-channel cases, adjustments to all six degrees of freedom (X, Y, Z, θx, θy, and θz) may be required (Figure 2). Effective fiber alignment requires the adjustment of several important motion parameters using a precision motion control device and an effective search algorithm suitable for a given application.
Motion degrees of freedom
Figure 2. Motion degrees of freedom.

Key Motion Parameters for Fiber Alignement

When using motion control systems for fiber alignment, the motion parameters considered for each axis critically affect the alignment process. The following are the primary parameters for consideration when selecting a motion controller for the location of peak power in fiber alignment procedures:

  • Minimum Incremental Motion (MIM) - This is the smallest increment of motion that a device can consistently and reliably deliver. It should not be confused with resolution, which is based on the smallest controller display value or smallest encoder increment. Rather, MIM is the actual physical performance of the controller that enables adjustment of the fiber position while searching for the position at which peak power is achieved. MIM of a motion controller can range from 100 nm to 1 nm. While a smaller MIM may align the fiber closest to the maximum peak power, this ability is achieved at significant costs in terms of alignment speed and power increments. XMS stages (Figure 3) are designed with optimized MIM and speed characteristics. They are capable of 1 nm MIM and 300 mm/s speed, making them ideal stages for alignment applications.
1 nm MIM of an XMS linear stage, XMS50-S linear motor stage
Figure 3. 1 nm MIM of an XMS linear stage (left); XMS50-S linear motor stage (right).
  • Repeatability - The repeatability parameter defines a motion control systemÕs ability to repeatably position. It can be unidirectional (always approaching the target position from the same direction) or bidirectional (approaching the target position from either direction). Bi-directional repeatability typically ranges from 1 µm to a few nm in fiber alignment systems. This parameter is important for quickly finding the peak power location for similar device designs. The XMS stage shown in Figure 3 has 80 nm bi-directional repeatability.
  • Position stability - Position stability is a measure of the motion system's capability to stay at a position within a defined window of time and error. Aligning fibers for assembly steps such as bonding relies on the positional stability of the fibers after the peak power has been located. Position stability requirements can range from 0.5 µm to a few microns. Figure 4 shows the step and settle performance of an MKS stage 250 ms after being moved. This stage exhibits less than 20 nm variation in position stability after settling.
Step, settle, and stability at position; XMS50-S linear motor stage (right)
Figure 4. Step, settle, and stability at position.
  • Other Motion Parameters - Other parameters that influence the effectiveness of a motion control system include: axis alignment, location of the gimbal point, system stiffness, pitch/yaw, thermal considerations, fixture design, Abbe error, etc. Details regarding the fundamentals of motion control can be found in Chapter 1, Section III.F.1. MKS also has an available metrology primer that provides further information on this topic.

Search Algorithms

In addition to understanding critical motion parameters, efficient fiber alignment requires the selection of a positional search algorithm appropriate to the application and to the step in the alignment procedure. Specific search algorithms are available for finding the first light, i.e., the periphery of a light beam, after which different algorithms that are faster and more precise are used to find the peak power location. The choice of the second algorithm depends on whether the beam has a Gaussian distribution or top hat profile with multiple peaks. Some algorithms can be used to profile both types of beams and can also be used in parallel. First light search methods include:

  • Raster scan - This is the simplest search method. It scans along one axis and indexes by a certain distance along another axis, then repeats the cycle. It is one of the quickest methods for finding the first light of the beam. The concept is shown in Figure 1 where the green line shows how the raster proceeds.
  • Spiral scan - Figure 5 illustrates this first light method, which tracks in the general area of the beam using a spiral motion by synchronizing the motion of the X and Y axes.
Spiral search for first light
Figure 5. Spiral search for first light (image used with permission of GBC&S Consulting Services).

After first light has been located, other algorithms are available that can find the peak power location:

  • Hill climb - This is a 2D search method based on finding the highest power within a certain path (Figure 6). The direction of the climb favors the location of the higher power region. The hill climb method is most effective when the beam has a Gaussian profile and when power quickly increases. The hill climb method, by itself, is not effective in finding peak power with flat beam profiles.
Hill climb algorithm
Figure 6. Hill climb algorithm.
  • Centroid - The centroid search (Figure 7) moves along one axis and finds the peak, then from that peak, moves along the second axis to get to the final peak. Centroid is useful with top-hat or multi-peak profiles.
Centroid search algorithm
Figure 7. Centroid search algorithm.
  • Dichotomy - Starting with large increments, a dichotomy search (Figure 8) initially searches one axis at a time until a peak is identified. The search then returns to the first peak. Within this first peak, another search cycle is performed using finer steps to find the maximum peak location.
Dichotomy search algorithm (image used with permission of GBC&S Consulting Services)
Figure 8. Dichotomy search algorithm (image used with permission of GBC&S Consulting Services).

Figure 8. Dichotomy search algorithm (image used with permission of GBC&S Consulting Services).

  • Other algorithms - There are many other algorithms that can be used either alone or in combination. A user should experiment with different combinations to find the most effective method for a specific beam profile or set of device characteristics.

MKS Semiconductor Handbook Cover

For additional insights into photonics topics like this, download our free MKS Instruments Handbook: Principles & Applications in Photonics Technologies

Request a Handbook