The wave-particle duality of light is an interesting contradiction for academics, but for designers of lighting systems, there are many practical, head-scratching contradictions and misconceptions related to light propagation in multi-mode step index fiber that can and do lead designers astray. Understanding and appreciating some of these nuances can ensure an optimal end design.
For example, many people believe that fiber collimates light, and some Ph.Ds will go to their grave convinced that they can preserve the angle of light in a fiber, but that’s simply not the case.
In theory, yes, a perfectly collimated beam with no divergence will not disperse with distance. In the real world, outside influences interfere with theory, causing light waves to change their propagation angle as they travel in a direction. This makes it almost impossible to have a perfectly collimated beam.
Even if light energy could be reasonably collimated, travel through multi-mode step index fiber will not maintain the collimation. In fact, impurities in the fiber and bends in the fiber material will disrupt any tendency toward collimation that may already exist as the light enters the fiber. In practice, over any reasonable length, a considerable number of fiber entry angles less than the acceptance angle, will emit at the full acceptance angle; the incoming beam will spread to fill the NA.
Speaking of light entering a fiber, many believe that using a low numerical aperture (NA) fiber will collect and focus light from a source better than a wider NA fiber. Lower NA fiber will only accept light within its narrow acceptance angle (which is twice the critical angle) and propagate only that light. Calculate your acceptance angle here.
The light emerging at the other end will be emitted at the narrow acceptance angle of the fiber, so it will seem more collimated, but the projected energy will be much less than what is projected by the same size fiber with a wider NA. To do both, collect and focus (or collimate), the majority of light from a source requires the use of high NA fiber and lenses on the output side to change the emitting angle and focus the beam..
Even if light hits the fiber’s core surface within the acceptance angle, it is still not “home free;” what is measured at the other end isn’t exactly all of the incident light. Designers need to account for Fresnel losses. These losses occur due to the differing refractive indices between air and the fiber, and cause some light to be reflected, albeit a small portion, about 4% at each end (for a single multimode step index fiber with a perfectly polished surface). Still, in critical applications, the small losses add up.
Another major contributing loss factor is Cladding effect. Designers sometimes believe the entire surface of a fiber collects light. Because a fiber is made up of a core glass and a clad glass only the core will transmit light. Cladding glass is not designed to propagate light waves; it’s only purpose is to maintain TIR. Cladding is not intended to transmit light. This cladding loss is equivalent to the total area of the fiber, less the area of the core. It could range from 5 to 25% (more or less) depending on cladding thickness. Typical loss is about 17% for common multimode step index fiber. Cladding thickness does effect core attenuation for specific wavelengths, and it also impacts a fiber’s life when exposed to heat steam and chemical corrosion—but that’s a topic for a different article.
To help offset losses and capture more light, it’s often assumed that we can cheat the laws of physics using an optical taper to collect and ‘focus’ light. Many believe tapers change the NA of the class system. Tapers do not change the NA; they do change the fiber’s acceptance angle at one end because the walls of a taper are not parallel. In general, the acceptance angle will be ‘normal’ at the large end, and greater at the smaller, tapered end.
Here’s the problem: while the smaller end will collect higher angle light, it collects less of it because the area is reduced. When used as a collection optic, tapers have no ability to increase light collection efficiency. That said, tapers have their place, especially for dental and industrial curing applications when used as an emitter to spread light as it exits the light guide.
Sometimes theory doesn’t match reality because, in spite of how hard we try, we can’t control outside influences to theoretical limits, and that’s very much the case with fiber optics and light propagation. Manage the disconnects by working with folks who have been hands on with fiber optic technology for decades and know how to weave reality in your favor.
Fiberoptics Technology Inc.
1 Quassett Road
Pomfret, CT 06258