Hollow-core fibers face a fundamental trade-off. A wide core couples light in efficiently — more of the input beam enters the fiber. A narrow core converts light efficiently — the confined mode has higher intensity, driving stronger nonlinear interactions. You cannot have both. The input wants the core wide; the physics wants it narrow.
The standard response is compromise: choose an intermediate diameter that sacrifices some of each. Every uniform fiber sits on this trade-off curve, and moving along the curve means gaining on one axis while losing on the other.
A tapered fiber dissolves the trade-off by making the constraint spatially varying. The input end is wide, for efficient coupling. The core progressively narrows along the fiber's length, increasing the intensity where nonlinear conversion occurs. The input and the physics each get what they want, in different locations.
The result: twofold enhancement of vacuum ultraviolet generation at 148.38 nm, for nuclear clock spectroscopy of thorium-229. The tapered geometry outperforms any uniform fiber on the trade-off curve because it is not on the trade-off curve. It left the curve by replacing a global constraint with a local gradient.
The principle generalizes. Many engineering trade-offs exist because the design is spatially uniform. Making one parameter (diameter, temperature, composition, porosity) vary smoothly along a coordinate can transform a fundamental limitation into a design parameter. The key requirement is that the competing objectives depend on position — that what matters at the input is different from what matters at the output. If so, a gradient can serve both without compromise.
Chemical reactors use temperature gradients for this reason. Gradient-index optics use refractive index gradients. The tapered fiber adds geometric gradients to the toolkit. In each case, the move is the same: recognize that the “fundamental” trade-off was actually a consequence of uniformity, and break the uniformity.
The trade-off was never in the physics. It was in the design constraint.