Broadband near-IR emission in glass co-doped with rare earth ions

Amplified spontaneous emission (ASE) sources are based on rare earth (RE) doped optical fibers and exhibit broadband emission, high output power, and good spatial coherence. In contrast to fiber lasers, they are characterized by low coherence and high spectral power density over their whole emission range. These unique properties make fiber ASE sources suitable for use in tunable radiation sources.

Superluminescent solid-state sources of radiation are frequently used in fiber optic communication, spectroscopy, optical sensors, fiber optic gyroscopes, and medicine. Such sources are attractive in offering the possibility of constructing rangefinders and laser radars in the ‘eye-safe’ region above 1.5μm. Several approaches have been explored with a view to broadband emission, such as doping glass with metal ions (for example, bismuth), generating supercontinuum light in photonic crystal fiber, and co-doping glass with lanthanide ions.^1–3 Among the lanthanides, erbium, thulium, and holmium ions are the most commonly used to construct co-doped sources of near-IR radiation.^4–6 Suitable combinations of doping RE elements make it possible to create broadband sources by exploiting their absorption and luminescence properties using energy transfer mechanisms between RE ions so that the emitted radiation is a superposition of their optical transitions.

Glass composition is key to achieving both practical and wide gain with a stable level of emission across the spectral range. The most frequently used RE host for fiber lasers is silicon dioxide (SiO₂), which is characterized by high optical transparency, mechanical resistance, and thermal stability. Its main drawback—beside the low solubility of RE ions—is high phonon energy (1100cm⁻¹), which leads to non-radiative transitions and consequently to inefficient energy transfer (ET) between RE ions. Thus, there is an eager search for new glasses with suitable mechanical properties and relatively low phonon energy for high energy transfer between lanthanide ions.

We have considered one possible solution, which is to refine the glass composition to control the vibrational frequency of phonons (collective excitations of atoms). Combining glass-forming elements with different phonon energies increases the physical separation of RE ions, which effectively broadens the spontaneous emission spectrum.⁷ We have suggested that the high content of low-phonon glass-forming elements results in RE ions being selectively surrounded by a phase rich in antimony oxide, minimizing the influence of high phonon energies from germanium oxide. Simultaneously, the presence of strong bonds in the glassy matrix ensures good thermal stability and mechanical properties, as required for optical waveguides.

We have used an antimony-germanate glass co-doped with erbium/thulium ions (Er³⁺/Tm³⁺) with broadband emission (full width at half-maximum, 420nm) resulting from the superposition of radiative transitions in erbium ions (centered at 1535nm) and thulium ions (centered at 1800nm). This system maximizes ET efficiency so that it can occur by radiative and non-radiative processes. The interaction between donors (Er³⁺) and acceptors (Tm³⁺) occurs at nearly equal energies between ground and excited states. Therefore, non-radiative quasi-resonant energy transfer is dominant.

To investigate ET mechanisms, we prepared a set of glass samples with constant (1mol%) concentration of Er³⁺ and variable (0.1, 0.25, 0.5, 0.75, and 1mol%) concentration of Tm³⁺ by adding erbium oxide (Er₂O₃) and thulium oxide (Tm₂O₃) to the glass. The high molar content of Er³⁺ enables strong absorption of direct pumping at 976nm. Increasing the Tm³⁺ ions reduces the 1535nm luminescence, confirming efficient quasi-resonant Er³⁺ → Tm³⁺ energy transfer. At the same time, increasing the concentration of Tm³⁺ above 0.5mol% quenches the 1800nm luminescence band, corresponding to the ³F₄ → ³H₆ (Tm³⁺) transition (see Figure 1).

Figure 1. Luminescence spectra of fabricated glasses for different molar ratios of erbium/thulium ions (Er³⁺/Tm³⁺), given as the ratio of erbium oxide (Er₂O₃) and thulium oxide (Tm₂O₃). Inset: The maximum of emission intensity of electronic transitions ⁴I_13/2 → ⁴I_15/2 (Er³⁺) and ³F₄ → ³H₆ (Tm³⁺) as a function of Tm³⁺ content. a.u.: Arbitrary units.

Optimizing the efficiency of the quasi-resonant energy transfer between Er³⁺ and Tm³⁺ was based on Förster-Dexter (semi-empirical) and experimental (luminescence measurement) methods. The first is a typical approach where the absorption of acceptor ions and emission of donor ions overlap. In the second method, we determined the efficiency of energy transfer between donor and acceptor by analyzing changes in the emission intensity at 1535nm in Er³⁺/Tm³⁺-doped glasses as a function of Tm₂O₃ concentration. Comparing both methods showed that controlling acceptor content plays an important role in optimizing ET (see Figure 2).

Figure 2. Comparison of energy transfer (ET) efficiency calculated with Förster-Dexter theory and luminescent measurement.

Figure 3. Comparison of normalized luminescence spectra of Er³⁺/Tm³⁺(red line), Er³⁺ (dotted line), and Tm³⁺ (dashed line) doped antimony-germanate glass. λ: Wavelength. FWHM: Full width at half-maximum.

Figure 3 shows a normalized luminescence spectra of Er³⁺/Tm³⁺, Er³⁺, and Tm³⁺ doped antimony-germanate glasses where, due to superposition of both luminescence bands, emission at approx. 420nm spectral width was obtained.

In summary, we have achieved broadband emission by optimizing the quantity of active dopants. To obtain efficient ET and broadband emission from Er³⁺ and Tm³⁺, 1mol% Er₂O₃ and 0.25mol% Tm₂O₃ is sufficient. Such glasses are suitable for fabricating broadband near-IR waveguide radiation sources, and our next step is to realize these.