The pulse energy that a fiber laser can make scales with the transverse dimension (core diameter) of the fiber. Unfortunately, we can’t just make fibers larger and larger, because large fibers guide multiple transverse modes. Multimode fibers typically generate speckled, low-quality beams that cannot be focused to a small spot (which is required for most applications of ultrafast lasers), and the transverse modes also tend to make pulses with poor quality in the temporal domain.

There are methods to generate clean beams from multimode fiber (see beam cleaning and wavefrontshaping approaches), but they’re complicated and don’t work well enough to justify in real-world use. There are also ways to make large-core, fiber-like rigid waveguides that only guide a single mode, but these are expensive, not scalable, and sacrifice a lot of the practical benefits of fiber.

I recently developed a new amplifier platform that enables single-mode (perfect beam) operation of extremely large-core, highly-multimode fiber (read about it here). The key idea is to use regenerative amplification. Rather than sending a pulse into a fiber and letting it come out the other end, like typical fiber amplifiers, in a regenerative amplifier the fiber is placed within an optical cavity that contains an optical switch. The switch traps the pulse in the cavity so that it undergoes many round trips through the amplifier before being dumped out. Due to some complex details of how the spatial modes behave within the cavity, the fundamental mode ‘automatically’ emerges, and the laser makes a perfect beam despite the highly-multimode fiber. This is a simple, robust solution to a complex problem, and it comes with an important practical benefit: we show that regenerative fiber amplifiers have much, much higher gain than is available in conventional, single-pass amplifiers. This makes the system– a ~60 dB single amplifier– more compact and inexpensive than typical amplifiers, which require a cascade of individual amplification stages (~20 dB each). The performance of this first-iteration already rivals the highest-performance ultrafast fiber amplifiers in the world today, in a much smaller, simpler package, and we’re optimistic that this idea will lead to order-of-magnitude improvements.