Fibers can be arranged in what are called one-dimensional (1D) and two dimensional (2D) arrays.

A 1D array is a single line of fibers, and a 2D array is a grid of fibers in both X and Y dimensions. In both cases, they are specified by the spacing between fibers. In the case of 1D array, there is a single defining dimension denoting the center-to-center position of the cores of the fibers.

For 2D arrays, there are two dimensions, X and Y, that define the core positions. One-dimensional arrays can be manufactured by several different techniques. The manufacturing technique will define the position accuracy of the fiber cores.

Fibers arranged in a machined slit (see Fig. below.) will be defined by the fiber diameter tolerance and the machined opening size and tolerance. This type of array is common for spectrometer slits. If more specific center positions are necessary, then machined holes can be drilled to position each fiber.

Machining tolerances again come into play, so one can expect positional accuracy of ±25–40 mm.

For the highest position accuracy, A v-groove is used to hold the fibers V-grooves can be made by grinding individual grooves into a glass substrate. A wide variety of materials, such as fused quartz, BK7, HK9L and other glasses, can be chosen to match other parameters of the assembly. This type of v-groove can be made to a very precise center to center tolerance of 1 mm or less. Glass v-grooves are usually specified to hold the fiber cores to ±1–2-mm positional accuracy after assembly.

Another technique for manufacturing v-grooves is by photolithography and KOH etching in silicon wafers. Photolithography is capable of submicron features in silicon, so this is the most accurate way to make a v-groove. Typical fiber positioning in an etched silicon wafer will attain ±0.5 mm or less in core-to-core measurements. Fiber position is dependent on factors other than just the v-groove substrate. Fiber tolerance, assembly technique, and epoxy choice play a role in how accurately the fibers can be placed in the termination.

Two-dimensional arrays range from a few fibers to thousands of fibers, depending on the applications. 2D arrays are used in optical switching and in sensing applications where spatial optical data is necessary, such as DNA sequencing, astronomy, and nuclear research.

Two-dimensional arrays can be made by similar methods as 1D arrays. A machined opening can hold multiple rows and columns of fibers. The fibers can be packed into this opening in two ways; a close (hexagon) packing or a square packing arrangement (see Fig. below). These two types of arrays have many uses such as illumination and specialty spectroscopy applications, but the center-to-center tolerance is highly dependent on fiber tolerance and traditional machining tolerances.

The most accurate way to place fibers in a 2D array is by way of individual holes for each fiber. One such method creates holes in a glass substrate by a process that uses light to alter the structure of the glass (using a femtosecond laser) to enable selective etching to remove the modified material. Sub-micron resolution has been attained with this method by Femto print SA. The most common method of creating holes for 2D arrays is by photolithography and a KOH etching process. This is the same process that was developed to create integrated circuits. Fiber-to-fiber positioning of ±1–2 mm is regularly achieved with this method. This is a cost-effective way to create multiple die on a single wafer. The desired hole pattern is created with a photolithography mask, the wafer is etched, and then the individual die are diced out of the wafer.

The highest-precision arrays are made with single-mode (SM) fibers and polarization-maintaining (PM) fibers. Arrays made with multimode fibers are dependent on the fiber tolerance which can be as much as 1–2 percent of its diameter, while SM fibers have a tolerance of as little as ±1 mm or even submicron levels. See Fig. below for examples of a precision SM array and a PM array with the stress rods in alignment.