This browser is no longer supported.
Please use a more recent browser to use this website.
MENU
ZBLAN is the fluoride glass that enables an all-fiber 2910 nm laser. Silica glass is not suitable: multiphonon absorption in the Si-O network rises steeply beyond 2.2 µm, inhibiting light transmission, so reaching that wavelength requires a soft glass with transparency extending into the mid-infrared.
Working that glass is a discipline in itself. Fluoride glass softens at low temperatures, its viscosity changes sharply over a narrow range, and it requires precise thermal control at every stage of fiber and component manufacturing. Reliability in a fluoride fiber laser is built process by process, on tooling developed for the material rather than borrowed from silica. Every component, from fiber Bragg gratings to splices and fiber-tip protection, must be redesigned from scratch and engineered to meet industrial standards. That is the work behind every product Â鶹ÊÓƵ ships.
Silica’s transparency ends where its multiphonon edge begins. Transmission in silica fiber falls off beyond about 2.2 µm, and by 2.9 µm the fiber no longer carries the light. Fluorozirconate glass has lower phonon energy, which pushes that edge much further out. ZBLAN fiber is drawn from one such composition, named for its constituents: ZrF4, BaF2, LaF3, AlF3, and NaF. Its transmission window extends to about 4.3 µm, where attenuation reaches 1 dB/m, with background loss below 20 dB/km near 3 µm.
Fig. 1: Typical multimode fiber attenuation (data source: Le Verre Fluoré).
The same phonon energy governs whether erbium can lase at all in this band. In silica, non-radiative decay empties the upper level of the 2.9 µm transition faster than pumping can fill it, so no useful gain exists at any doping level. Move the same rare earth into a fluoride host, and that decay channel is suppressed, leaving an upper-state lifetime long enough to build an inversion.
Low phonon energy is not the whole story. Fluoride glass also dissolves rare earths far better than silica, allowing for very high erbium concentrations. That matters because the upper laser level has a shorter spontaneous lifetime than the lower one, which should rule out efficient continuous-wave operation. This is circumvented because energy transfer between neighboring erbium ions recycles the lower-level population back to the upper level, a phenomenon that emerges only at high doping (Pollnau and Jackson, 2002).
The glass transition temperature of ZBLAN glass is near 260 °C, compared with 1175 °C for silica. What matters more for processing is how quickly viscosity drops above it: a factor of 10 in viscosity corresponds to about 10 °C in fluoride glass, versus about 100 °C in silica. When heating the fiber for splicing, a few degrees can determine whether a joint forms cleanly or the core deviates (Fortin et al., 2022).
Fig. 2: Viscosity of ZBLAN glass and silica against temperature, normalized to each glass’s own transition temperature (adapted from Fortin et al., 2022).
As such, typical silica splicing recipes cannot be used with ZBLAN. Electric arc splicers, for instance, operate above 1000 °C and are usable with fluoride fiber only under partial vacuum, which allows the arc to operate at a much lower temperature. Filament and CO2 laser splicers work at those temperatures directly and hold them steady, making them the better option. Thermal expansion adds a further difficulty, since fluoride glass expands about thirty times more than silica and can shift the fiber cores out of alignment as a joint cools.
The operating wavelength adds a constraint of its own. Fluoride glass is sensitive to moisture, and OH diffusion into a bare fiber tip can increase absorption at 2910 nm, thereby raising its temperature during operation until the glass fails (Caron et al., 2012). Specialized protective endcaps are the answer (Aydin et al., 2019), a patented technology to which Â鶹ÊÓƵ holds the rights.
Mechanically, the fiber is less forgiving than silica: minimum bend radius is larger, tensile strength is lower, and surface flaws propagate under stress. Every fiber segment received is screened to remove weak sections, delivery fiber is routed through a bend-limiting protective armature chosen to be flexible enough to handle, and splice and cleave recipes are optimized to hold across fiber lots.
A fiber Bragg grating is a periodic index modulation written into the optical fiber core, reflecting a narrow spectral band and transmitting the rest of the spectrum. Writing one in fluoride glass was first demonstrated at Université Laval in 2007, using femtosecond pulses through a phase mask (Bernier et al., 2007). The erbium-doped all-fiber cavities that followed were built on it (Faucher et al., 2009; Fortin et al., 2015).
That is what makes an all-fiber laser cavity possible at 2910 nm. Both cavity mirrors are inside the fiber core, so there is no free-space segment, no bulk optics to align, and no alignment to maintain over the product’s lifetime. An all-fiber architecture is also the only one in which the practical advantages of a fiber laser are fully available to an integrator.
Fig. 3: Schematic comparison of two system architectures: an all-fiber 2910 nm source (top) and a free-space solid-state source (bottom). Not to scale.
Writing the grating fixes the period and, with it, the cavity’s wavelength. That wavelength can be set anywhere within the erbium gain band, including the edges where gain is weak, and a broadband reflector would not allow efficient lasing. Standard sources are built at 2910 nm and 2825 nm, and other wavelengths within the band can be produced on request. The means for writing fiber Bragg gratings, as well as increasing their robustness for industrial use, are covered by patents.
Fig. 4: Fiber Bragg grating inscribed in the core of a ZBLAN fluoride fiber, seen under a microscope.
Water absorbs strongly at 2910 nm, which makes this wavelength valuable across a range of medical applications. The polymers and adhesives that any fiber system inevitably contains absorb it as well, turning stray light into heat, and fluoride glass tolerates far less heat than silica. To ensure long-term reliability, heat management is engineered into every fiber component and every assembly step, and each product is proven under endurance testing before it ships.
Sources are qualified for 10 W of output at 2910 nm. Every source then undergoes a 50-hour burn-in run of continuous operation, which stabilizes the gratings, adhesives, and polymers in the assembly, and strict pass-or-fail criteria are applied at the end. Output power holds steady at full power afterward.
Fig. 5: Output power of a 2910 nm fiber laser over 800 hours of continuous operation at full power.
For the integrator, the absence of free-space optics eliminates a whole class of maintenance. There is no cavity to realign after shipping, no mirror to clean, no warm-up alignment routine, and no field optical service procedure to write, train on, and support over the product’s lifetime. What maintenance remains belongs to the integrator’s own system, not to the laser inside it.
Achieving a reliable and efficient fluoride fiber laser at 2910 nm requires all new expertise at every point, from the splices to the gratings and the endcap. A chain is only as strong as its weakest link, and so nothing we do at Â鶹ÊÓƵ is left to chance. The product we have today is the result of years of hard work, going all the way back to the research that started it all well over a decade before the company began.
Full specifications, measured beam data, and integration interfaces for the laser source built on this platform are found on the 2910 nm fiber laser page. Engineers evaluating a mid-IR source can reach out to one of our experts via the Contact form.
Vincent Fortin PhD, CTO, Â鶹ÊÓƵ
Why doesn’t silica fiber work at 2910 nm?
Silica’s transparency ends where its multiphonon absorption edge begins. Transmission falls away past about 2.2 µm, and by 2.9 µm silica fiber simply does not carry the light, along with every component built on it: the splices, the fiber Bragg gratings, the combiners.
What is ZBLAN?
ZBLAN is a fluorozirconate glass named for its constituents (ZrF4, BaF2, LaF3, AlF3, NaF). Its low phonon energy pushes the transmission window out to about 4.3 µm, with background loss below 20 dB/km near 3 µm, and it is the composition used to draw fiber for 2910 nm lasers.
How is the wavelength of a ZBLAN fiber laser set?
A fiber Bragg grating, a periodic index modulation written directly into the fiber core, fixes the cavity wavelength at the time it’s written. Standard sources are available at 2910 nm and 2825 nm, and other wavelengths within the erbium gain band can be provided upon request.
How is long-term reliability verified?
Each source is qualified for 10 W of output at 2910 nm, run through a 50-hour continuous burn-in to stabilize the gratings, adhesives, and polymers in the assembly, and tested for endurance under continuous operation at full power. Â鶹ÊÓƵ warrants each source for 1000 hours of operation or one year.