Open Access
Add to BibSonomy
Abstract
Erbium lanthanum titanate glasses were prepared by levitation melting for the spectroscopic study of ways to promote the mid-infrared fluorescence. Two series of heavily erbium doped glasses (15 wt%) were prepared with the addition of either Pr3+ or Nd3+ in amounts relative to Er3+ of 0.05, 0.1, and 0.2. Both ions quench the lower Er3+ laser level with the Pr3+ doing so more rapidly. Although high co-dopant concentrations result in higher energy transfer, as clearly evidenced in upconversion and downconversion fluorescence measurements, the mid-infrared lifetime also suffers a reduction and, therefore, a balance must be struck in the co-dopant concentration. Lifetime and spectral measurements indicate that, at a fixed relative co-dopant amount, Pr3+ is more effective than Nd3+ at removing the bottleneck of the Er3+ 4I13/2 level. Moreover, consideration of the lifetimes alongside the absorption data of the individual ions indicates that despite the large absorption cross-section of Nd3+ at 808 nm, the concentration needed to yield more absorbed power than utilizing direct 976 nm excitation of Er3+ results in unfavorable lifetimes of the mid-infrared transition. In the end, Pr3+ prevails as the superior co-dopant in terms of the effects on fluorescence lifetimes as well as potential laser system design considerations. In a unique self-doping approach, a reducing melt atmosphere of Ar instead of O2 creates a small fraction of Ti3+. In 5Er2O3-12La2O3-83TiO2 glass, the presence of Ti3+ quenches the 4I13/2 emission about 2.6 times more than the 4I11/2 when lifetimes are compared to an O2 melt environment. As an additional means of increasing the mid-infrared emission, the effect of temperature on the mid- and near- infrared lifetimes of a lightly doped lanthanum titanate composition is investigated between 77-300 K. The mid-infrared lifetime increases by ∼30% while the near-infrared lifetime increases by ∼10%, which suggests in addition to co-doping, active cooling of the gain media will further enhance performance.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Mid-infrared (MIR) lasers utilizing Er$^{3+}$ doped glasses and crystals receive a significant amount of attention for important applications such as in medical and defense settings [1]. Searching for new materials suitable for the class of lasers based on Er$^{3+}$ requires careful consideration of the relative lifetimes of the metastable levels $^4I_{11/2}$ and $^4I_{13/2}$ (see Fig. 1). The $^4I_{13/2}$ level lifetime is longer than that of $^4I_{11/2}$. This makes laser operation of the $^4I_{11/2} \rightarrow ^4I_{13/2}$ transition self-terminating. Oxide materials often possess superior physical, thermal, and mechanical properties compared to non-oxide counterparts, and so working around this bottleneck in these materials is worthwhile. Three common approaches to this are (1) heavy doping to increase energy transfer between erbium ions [2] (2) co-doping with ions that deactivate the lower laser level [3] and (3) making both mid- and near-infrared (NIR) transitions lase simultaneously [4].
We have been extensively studying the optical properties of undoped [5,6] and doped (Er$^{3+}$ [6], Dy$^{3+}$ [6], Yb$^{3+}$ [7]) lanthanum titanate glasses. Since the titanate glass backbone is comprised of an octahedral TiO$_n$ network (n=5-6) - a rare occurrence for oxide glasses [8,9] - the phonon energy ($\sim$730 cm$^{-1}$) is low enough to allow good mid-infrared transmission and favorable spectroscopic properties. Although the lifetime of the 3-micron emission from Dy$^{3+}$ was too short to warrant further investigation, Er$^{3+}$ displayed mid-infrared lifetimes of $\sim$300 $\mu$s. These results suggest additional study of the lanthanum titanate glass host as a potential gain media is valuable. However, the NIR lifetime in singly doped erbium lanthanum titanate glasses is about 10 times larger than the MIR. In this work, we investigate methods of enhancing the relative lifetime of the MIR emission, primarily through co-doping.
Co-doping with Pr$^{3+}$ was the first route to quenching the $^4I_{13/2}$ level in crystalline yttrium lithium fluoride and heavy metal fluoride glasses [10–12]. Other co-doping schemes that have been suggested include Nd$^{3+}$ [13–15], Ho$^{3+}$ [16], and Eu$^{3+}$ [17]. Nd$^{3+}$ appears numerous times in the literature, touted for the ability to sensitize absorption via the strong 808 nm absorption band of Nd$^{3+}$ while simultaneously deactivating the lower laser level. Studies on Ho$^{3+}$ and Eu$^{3+}$ are more scarce and not as promising. Lv et al. compared Eu/Er and Ho/Er co-doping in CaYAlO$_4$ for MIR erbium emission and suggested Ho$^{3+}$ as a better co-dopant for this goal [17]. Similarily, Xia studied Ho/Er and Pr/Er in SrGdGa$_3$O$_7$ and concluded Pr$^{3+}$ is superior for mid-infrared interests. Most, if not all, reports of MIR lasing in a co-doped Er:glass or Er:crystal laser are with Pr$^{3+}$ addition. Given this, and the relatively large number of spectroscopic reports suggesting the use of Nd$^{3+}$, we directly compare Pr/Er and Nd/Er co-doping here.
As an additional (somewhat lateral) contribution, temperature dependent lifetime measurements of the $^4I_{11/2}$ and $^4I_{13/2}$ levels are carried out using a singly doped glass of composition 0.5Er$_2$O$_3$-16.5La$_2$O$_3$-83TiO$_2$. This action is inspired by the work of Sanamayan who carried out dual-wavelength laser operation at cryogenic temperatures of Er:YAG with MIR and NIR outputs of 10 and 45 W, respectively [4]. In the accompanying spectroscopic measurements of the work, the emission cross-section of the mid-infrared $^4I_{11/2} \rightarrow ^4I_{13/2}$ transition grew by about a factor of 10 at 77 K relative to room temperature measurements [4]. Since here we aim to gain insight into methods by which the mid-infrared emission of Er$^{3+}$ in the titanate host can be fostered, low temperature lifetime measurements are warranted.
2. Materials and methods
Glasses were prepared using the levitation melting technique in oxygen gas [18]. To probe influence of melt atmosphere [19], one singly doped sample with 5% Er$_2$O$_3$ was melted in argon gas, which resulted in a dark colored glass. An undoped sample was melted in a stream of 5% CO with an argon balance, to further promote the reduction of Ti$^{4+}$ to yield a good electron paramagnetic resonance (EPR) signal. Singly doped samples are from our previous investigation [6]. Co-doped samples followed the composition rule xNd$_2$O$_3$-5Er$_2$O$_3$-(12-x)La$_2$O$_3$-83TiO$_2$ or (x/3)Pr$_6$O$_{11}$-5Er$_2$O$_3$-(12-x)La$_2$O$_3$-83TiO$_2$ where in both cases $x$=0.25, 0.5, and 1. This yielded relative ratios of Nd (Pr) to Er of 0.05, 0.1, and 0.2.
Initially lifetime measurements on co-doped glasses were also carried out at room temperature with a mechanically chopped 976 nm laser, but the very short lifetimes encountered became on the order of the pulse width ($\sim$25 $\mu$s) and thus the decay behavior of the excited erbium ions could not be discerned unambiguously. To resolve this, a frequency doubled Nd:YAG laser was used as the excitation source. The pulse duration was 550 ps with a repetition rate of 500 Hz. The trends observed were the same with either source.
Static spectral measurements were made using various spectrometers for UV-VIS-NIR (Ocean Optics HR4000), NIR (Yokogawa AQ6370D), and MIR (Redstone OSA305). Visible up-conversion and NIR/MIR down-conversion were measured using 976 nm excitation. Absorption measurements were taken via white light imaging using a stabilized tungsten halogen source and application of the Beer-Lambert law. Glasses following the composition rule 2RE$_2$O$_3$-15La$_2$O$_3$-83TiO$_2$ were prepared as spheres and then ground and polished to create approximately 1 mm thick discs for absorption measurements of RE$=$ Er, Nd, Pr. Spectra from the three different spectrometers were stitched together. No absorption was measured for Er$^{3+}$ or Nd$^{3+}$ above the 1700 nm cut off of the Yokogawa AQ6370D spectrometer using the Redstone OSA305 instrument.
Fluorescence lifetime measurements for the temperature-dependent study of a singly doped 0.5Er$_2$O$_3$-16.5La$_2$O$_3$-83TiO$_2$ glass disc were carried out in a liquid nitrogen cryostat equipped with an electrical heater. The sample was clamped to the copper cold finger. Excitation came from a mechanically chopped 976 nm or 1480 nm diode to directly excite the $^4I_{11/2}$ or $^4I_{13/2}$ level, respectively. The pulse width was approximately 25 $\mu$s with a 10 Hz repetition rate.
EPR data were collected on 17La$_2$O$_3$-83TiO$_2$ glasses using a Bruker EMX spectrometer and Xenon software. Glass beads, approximately 0.1 g in total mass, were loaded into 5 mm borosilicate tubes for X-band (9.44 GHz) EPR measurements at room temperature. Measurements were conducted using microwave powers of 51.68 mW, field modulation amplitudes of 4 Gauss, and signal averaging of 10 scans. Estimates for the Ti$^{3+}$ concentration in the 5%CO:95%Ar-processed glass were made by comparing the signal area (double integration of the full spectrum) with that of a glass containing a known concentration of Cu$^{2+}$.
3. Results and discussion
Figure 2 displays absorption spectra for lanthanum titanate glasses containing about 2 mol% of either Er$_2$O$_3$ (black), Nd$_2$O$_3$ (red), or nominally Pr$_2$O$_3$ (blue). Plots are presented over 388-2200 nm, corresponding to the UV edge of the base lanthanum titanate glass [6] and near the mid-infrared cutoff of the multimode fiber employed in the absorption measurements. The positions of the ground state absorption bands are, as expected, typical for Er$^{3+}$, Nd$^{3+}$, and Pr$^{3+}$ in glasses.
Fig. 2. Absorption spectra for lanthanum titanate glasses containing 2 mole percent RE$_2$O$_3$ for RE= Er (black), Nd (red), and Pr (blue). Absorption measurements were taken on 1 mm thick polished discs.
Near and mid-infrared emission spectra for 5Er$_2$O$_3$-12La$_2$O$_3$-83TiO$_2$ under 976 nm excitation is shown in Fig. 3, where all spectra are normalized to the 2.7 $\mu$m band maximum. The emission intensity of both the 1.5 and 2.7 $\mu$m transitions is comparable here. During our previous work, for glasses containing on the order of about a tenth of the erbium concentration the 1.5 $\mu$m transition intensity was significantly greater, by about a factor of 10. Quenching of the emission at $^4$I$_{13/2}$ was previously observed via lifetime measurements to be greater for erbium concentrations of several mole percent [6]. Excited state absorption or cross relaxation processes from the $^4$I$_{13/2}$ level might diminish the population, both of which can be reasonably expected at the moderately high concentration of 5 mol%. Reabsorption of $^4$I$_{13/2} \rightarrow ^4$I$_{15/2}$ is also expected to influence the apparent intensity, as a measurement on a smaller bead of 5% Er$_2$O$_3$ showed stronger intensity of the 1.5 $\mu$m band. To limit impact of size, the measurements shown in Fig. 3 are all sample spheres of $\sim$2 mm in diameter. Reabsorption of fluorescence and the (sometimes significant) impact on measured spectra has been studied in detail by others, for example Refs. [20,21]. Here, we restrict comparison to samples of nominally the same size and measured under identical circumstances.
Fig. 3. Near and mid-infrared fluorescence spectra of 5Er$_2$O$_3$-12La$_2$O$_3$-83TiO$_2$ (black), 0.05Nd$_2$O$_3$-5Er$_2$O$_3$-11.95La$_2$O$_3$-83TiO$_2$ (red), and 0.05Pr$_2$O$_3$-5Er$_2$O$_3$-11.95La$_2$O$_3$-83TiO$_2$ (blue) measured under 976 nm excitation.
The introduction of co-dopants at a 1:20 ratio relative to the erbium concentration is also presented in Fig. 3. The erbium concentration has been fixed, and Fig. 2 shows no overlapping absorption band from neodymium or praseodymium ions at 976 nm. Therefore, the absorbed power is the same and any spectral changes are due to excited erbium ions interacting with the added Nd or Pr atoms. Begining with the Nd$^{3+}$ doped glass, for the 0.05 Nd/Er spectra, the relative intensity of the 1.5 $\mu$m emission decreases. A weak but distinct feature appears with a maximum near 1.85 $\mu$m can be assigned to the $^4F_{3/2} \rightarrow ^4I_{15/2}$ transition of Nd$^{3+}$ [22,23]. The more common emissions from Nd$^{3+}$ were observed at 1.07 $\mu$m ($^4F_{3/2} \rightarrow ^4I_{11/2}$) and 1.35 $\mu$m ($^4F_{3/2} \rightarrow ^4I_{13/2}$) under 915 nm excitation, but were not detected under 976 nm excitation. Another feature of the 0.05Nd/Er spectrum, compared to the Er-only spectrum of Fig. 3, is the increased intensity of the long-wave tail beyond 4 $\mu$m. This is anticipated to belong to increased blackbody radiation from a higher sample temperature resulting from energy migration to the added Nd atoms followed by non-radiative decay events.
The spectrum for 0.05 Pr/Er shows only the mid-infrared emission from Er$^{3+}$ at 2.7 $\mu$m alongside a strong long wavelength tail. The stronger quenching of the 1.5 $\mu$m Er$^{3+}$ emission with praseodymium co-doping is understood readily by referring back to Fig. 2. Pr-doped lanthanum titanate has a very weak absorption band near 1000 nm ($^3H_4 \rightarrow ^1G_3$) and a strong absorption band spanning 1300-1700 nm ($^3H_4 \rightarrow ^3F_4, ^3F_3$). This leads to efficient energy transfer from the Er$^{3+}$ $^4I_{13/2}$ levels to Pr$^{3+}$. Once the Pr$^{3+}:$ $^3F_4, ^3F_3$ levels are populated, the decay back to the ground state is dominated by non-radiative decays due to the close proximity of the $^3F_2$ and lower levels. This route has been known for a while [12], and it is well exemplified here by the long wavelength tail in the 0.05 Pr/Er spectrum of Fig. 3, where in this work it appears the intensity of this feature is proportional to the non-radiative decays.
Fluorescence decay curves for glass co-doped with Pr and Nd are shown in Fig. 4. The mid-infrared decay for 0.05Pr:Er glass is nearly mono-exponential. Aside from this, all decay curves appear multi-exponential. To facilitate comparison, the expectation integral is used to determine a mean lifetime, $<\tau >$,
where $I(t)$ is the normalized fluorescence intensity. The results of the calculations are presented in Fig. 5(a) for Pr:Er glasses and Fig. 5(b) for Nd:Er glasses. A figure of merit of the change in lifetime can be represented by an energy transfer efficiency parameter, $\eta _{ET}$, where $\tau _{RE/Er}$ refers to the lifetime of a co-doped glass and $\tau _{Er}$ to the lifetime of an Er-only doped glass (see Fig. 5(c),(d)).Fig. 4. Mid infrared (black) and near infrared (red) decay curves under pulsed 532 nm excitation for Pr:Er ratios of 0.05 (a), 0.1 (b), and 0.2 (c) and for Nd:Er ratios of 0.05 (d), 0.1 (e), and 0.2 (f).
Fig. 5. Mid-infrared (black) and near infrared (red) lifetimes for Pr/Er co-doping (a) and Nd/Er co-doping (b). Mid-infrared (black) and near infrared (red) energy transfer coefficients for Pr/Er co-doping (c) and Nd/Er co-doping (d).
For both Pr and Nd, the trends of both lifetimes and energy transfer efficiency are strikingly similar in the compositions of this study. The measured lifetimes of the MIR and NIR Er$^{3+}$ transitions are presented in Table 1. At just 1:20 Pr:Er ratio, the lifetime of $^4I_{11/2}$ becomes longer than $^4I_{13/2}$ indicating excellent deactivation. At the Er$_2$O$_3$ levels and co-doping ratios used here, the $^4I_{11/2}$ lifetimes are below 100 $\mu$s which is rather undesirable. Some fine tuning in the optimal Er:Pr ratio and overall Er$_2$O$_3$ concentration is required for laser development, but this work here suggests this search should be conducted with Pr:Er ratios in the vicinity 0.01-0.05. Inspection of Fig. 5 suggests that for such low Pr content, the lifetime of the lower laser level is not fully quenched, but attainment of population inversion in $^4I_{11/2}$ could occur due to synergistic effects of cross-relaxation between Er$^{3+}$ ions and deactivation via Pr$^{3+}$. Slightly lower Pr$^{3+}$ concentration may also reduce the thermal load needed to be managed from high rates of non-radiative decays as evidenced by Fig. 3.
Table 1. Fluorescence lifetimes for 5%Er$_2$O$_3$ lanthanum titanate with co-dopant RE = Pr, Nd
Interest in Nd$^{3+}$ as a co-dopant lies in the utilization of the ion’s large cross-section at 808 nm to increase pump absorption. The absorption cross-sections ($\sigma _{ab}$) were calculated from the measured spectra in Fig. 2. For Nd$^{3+}$, $\sigma _{ab}(808\,\rm {nm})$=3.16 pm$^2$ whereas for Er$^{3+}$ $\sigma _{ab}(798\,\rm {nm})$=0.16 pm$^2$, giving roughly a factor of 20 difference. Assuming a 100% transfer of pump energy at 808 nm, the ratio of Nd:Er is required to be in excess of 1:20 in order for the system to benefit from the Nd$^{3+}$ presence. Since for Er$^{3+}$ in the lanthanum titanate glass, $\sigma _{ab}(976\,\rm {nm})$=0.38 pm$^2$, the quantity of Nd$^{3+}$ required to increase the pump absorption to be superior to simply pumping at 976 nm is greater than 1Nd:10Er. At these ratios, the energy transfer out of Er$^{3+}:\, ^4I_{11/2}$ exceeds 80% and lifetimes are below 100 $\mu$s, making this route undesirable for actual laser development. As a final note along this train of thought, pumping at 808 nm additionally yields a 6% larger quantum defect for lasing at 2750 nm compared to 976 nm pumping, in turn requiring more heat management and lower maximum efficiency (ignoring Er-Er transfer processes such as cross-relaxation). Considering these factors, Nd$^{3+}$ co-doping does not appear as promising as Pr$^{3+}$ for the system under study for mid-infrared laser applications.
To further probe the energy transfer and subsequent decay dynamics, the upconversion spectra under 976 nm excitation were recorded. The static fluorescence spectra are shown in Fig. 6 for fixed excitation power and therefore, as discussed above, a fixed absorbed power for 976 nm excitation. On a per atom basis, the impact from Pr and Nd are also highly similar, as was observed for down-conversion NIR and MIR emissions in Fig. 5(c),d. Figure 6 shows the three most commonly observed erbium upconversion transitions: $^4H_{11/2} \rightarrow ^4I_{15/2}$, $^4S_{3/2} \rightarrow ^4I_{15/2}$, and $^4F_{9/2} \rightarrow ^4I_{15/2}$. The intensity of the green band at $\sim$550 nm ($^4H_{11/2} + ^4S_{3/2} \rightarrow ^4I_{15/2}$) decreases more for Nd addition while the red band at $\sim$680 nm ($^4F_{9/2} \rightarrow ^4I_{15/2}$) decreases more for Pr addition. This suggests slightly different energy transfer from the higher lying Er levels to the different co-dopants, but similar magnitudes for a given co-dopant concentration. Also, the same trend as observed for downconversion fluorescence is observed wherein the energy transfer out of erbium increases with increasing co-dopant concentration.
Fig. 6. Upconversion fluorescence spectra of Nd or Pr co-doped erbium lanthanum titanate glasses under 976 nm excitation. The asterisks indicate stray light from the laboratory.
Excited state absorption is a common driver of upconversion fluorescence in Er$^{3+}$ doped materials. For ESA, the slope of the log-log plot of the emission intensity versus the excitation power should be approximately 2, representing a two-photon process. Since in Fig. 6 the emission from the $^4F_{9/2} \rightarrow ^4I_{15/2}$ is weak, as seen clearly by the proximity to the noise floor of the spectrometer for high co-dopant concentrations, the discussion here is limited to the $^4S_{3/2} \rightarrow ^4I_{15/2}$ transition by integrating the fluorescence intensity over 538-580 nm. A double logarithmic plot of the integrated spectral intensity versus the excitation power as measured for the $^4S_{3/2}$ transitions is shown in Fig. 7(a). At relatively low powers (Fig. 7(b)), the singly doped glass has a slope $\sim$1.9, indicating a two-photon process driving the population of the higher lying levels. For these same pump powers, the fluorescence intensity of the co-doped glasses was too low to be measured reliably with fixed measurement parameters.
Fig. 7. (a) Log-log plot of integrated area between 538-580 nm in the upconversion spectra under 976 nm excitation. Numerical values next to data point groups with linear fits indicates the slope for the fitted region, between either 1.8-2.1 or 2.3-2.6 on the $x$ axis. (b) Log-log plot of the same area in (a) but for singly doped glass and at lower excitation, revealing a slope near 2 indicative of excited state absorption. (c) Fluorescence decay curves $^4S_{3/2} \rightarrow ^4I_{15/2}$ under mechanically chopped 976 nm excitation. Note the mechanically chopped pulse is $\sim$25 $\mu$s wide.
At higher pump powers, the visible upconversion of all glasses were observable (Fig. 7(a)). Two trends are observed (1) decreasing slope, n, with increasing excitation power and (2) increasing slope with increasing co-dopant concentration. The phenomena of (1) is consistent with saturation of the level. The occurrence of (2) is likewise suggested to be related to the saturation after considering the lifetime of $^4S_{3/2} \rightarrow ^4I_{15/2}$ under 976 nm excitation (Fig. 7(c)). Although the pulse width quickly becomes comparable to the decays observed here, and so a high uncertainty can be inferred, qualitatively a clear decrease in the lifetime of the green emission is evident from the data. Since the saturation power is inversely proportional to the lifetime, the observed faster decays with greater co-dopant concentrations implies larger saturation powers for the co-doped glasses. This then explains the trend (2) phenomena.
Up until now, we have considered only the traditional rare-earth co-doping schemes. During the course of this work, the effect of the melt atmosphere was investigated. The composition 5Er$_2$O$_3$-12La$_2$O$_3$-83TiO$_2$ was melted under argon to create reducing conditions. For the glass prepared as so, under 976 nm excitation, the lifetime of $^4I_{13/2}$ decreases from 2.6 ms to 1.1 ms whereas the lifetime of $^4I_{11/2}$ only decreases from 220 $\mu$s to 170 $\mu$s - both relative to the same composition prepared under typical O$_2$ flow (see Fig. 8(a),b). To probe this phenomena, an undoped glass was prepared for EPR spectroscopy. A strongly reducing environment was used in this case, with 5% CO and an argon balance. This was selected to maximize the signal from the anticipated Ti$^{3+}$.
Fig. 8. Fluorescence decay curve in NIR (a) and MIR (b) for 5Er$_2$O$_3$-12La$_2$O$_3$-83TiO$_2$ glass prepared under oxidizing (O$_2$, black) and reducing (argon, red) conditions. (C) Normalized EPR signal of undoped lanthanum titanate glass prepared under oxidizing (O$_2$, black) and strongly reducing (5%CO:95%Ar, red) conditions
The EPR signals are plotted in Fig. 8(c), each normalized to 1.0 gram of glass and corrected for differences in the resonator tuning characteristic. The strong peak for the 5%CO:95%Ar glass is due to Ti$^{3+}$ and the weak peak in the O$_2$ glass is ferric iron (approximately 200 ppm Fe$^{3+}$), confirming the higher oxidation for this preparation route. Presumably, the iron is mostly ferrous in the 5%CO:95%Ar processed glass and thus not detected. Vertical expansion of the 5%CO:95%Ar glass does show the presence of a small amount of Fe$^{3+}$, on the order of 1 ppm. At any rate, the total Ti$^{3+}$ concentration for the 5%CO:95%Ar glass is small, $\sim$1230 ppm.
The EPR results for the strongly reducing atmosphere and the quenching and dark coloration of the erbium glass prepared under just Ar flow suggest the reduction of Ti$^{4+}$ to Ti$^{3+}$ is present in the latter case. This demonstrates a unique aspect of the transition metal oxide based host, that network forming species can self-dope the host and interact with active rare-earth species. Now, this only shows first observations and the compatibility with rare-earth co-doping is unclear. Nevertheless, the EPR measurements and the favorable reduction of the $^4I_{13/2}$ lifetime relative to $^4I_{11/2}$ in these measurements suggests two things. One, that the melt atmosphere has non-trivial influence on the titania oxidation state. And two, that this influence yields an additional adjustable parameter in the preparation of optically active media based on rare-earth titanate glasses.
Lastly, the temperature dependence of the MIR and NIR lifetimes are measured for a sample with 0.5 mol% Er$_2$O$_3$. The comparatively low concentration here, relative to the rest of this study, was opted for to minimize effects of NIR reabsorption in the experimental setup. The results are shown in Fig. 9. Figure 9(a),b show the lifetimes near room temperature as well as at liquid nitrogen temperature. In both cases, the eye easily sees that the lifetime is increased at low temperature, as expected. Figure 9(c),d show the measured lifetime values at different temperatures between 77 and 300 K. The decay curves here are much simpler than the co-doped glasses (Fig. 4) and so the NIR lifetimes here were obtained from mono-exponential fits while the MIR lifetime were obtained by bi-exponential fits. The plotted error bars represent the uncertainty from these procedures. At 300 K, the MIR and NIR lifetimes are 283 $\mu$s and 3342 $\mu$s, respectively. As the temperature decreases to 77 K, the lifetimes increase to 362 $\mu$s and 3733 $\mu$s. These measurements suggest that in addition to slight Pr$^{3+}$ co-doping, cryogenic cooling will help in the enhancement of MIR lasing from titanate glass and that - at a minimum - active cooling near room temperature will be required to avoid significant thermal quenching of the MIR emission.
Fig. 9. (a,b) Fluorescence decay curves at 77 K (blue) and 300 K (black) for $^4I_{11/2}$ (a) and $^4I_{13/2}$ (b). (c,d) Lifetime versus temperature for $^4I_{11/2}$ (c) and $^4I_{13/2}$ (d). The broken red lines are drawn to guide the eyes.
4. Conclusion
Lanthanum titanate glasses doped heavily with erbium and lightly co-doped with either Nd$^{3+}$ or Pr$^{3+}$ were fabricated via levitation melting. Fluorescence spectroscopy revealed substantial deactivation of the upconversion and lower laser level of Er$^{3+}$ with either ion and in all concentrations studied. Pr$^{3+}$ appears superior to Nd$^{3+}$ for deactivation processes since the lifetime of the MIR laser level becomes longer than that of the lower level with as a few as 1 Pr atom per 20 Er atoms. The benefits of employing Nd$^{3+}$ as a sensitization agent are challenged in this work. Despite the large absorption cross-section of Nd$^{3+}$ at 808 nm, pumping this level does not seem likely to increase absorbed pump power compared to directly pumping Er$^{3+}$ at 976 nm, due to the likelihood of encountering severe quenching of the MIR lifetime at concentrations needed to achieve this. Self-doping via partial reduction of Ti$^{4+}$ to Ti$^{3+}$ produces favorable lifetime changes as well, and this phenomena should be pursued with different Er concentrations, co-dopants, and melt atmosphere conditions in future work to investigate the viability of this approach. Temperature-dependent lifetime measurements showed the MIR lifetime receives a $\sim$30% enhancement at low temperatures, about a factor of three times larger than the enhancement of the NIR lifetime.
Funding
National Science Foundation (2217786); U.S. Department of Energy (DE-SC0018601); National Aeronautics and Space Administration (80NSSC19K1288); Air Force Research Laboratory (FA9451-22-2-0016).
Acknowledgments
BT gratefully acknowledges support in the form of a Postdoctoral Appointment under the supervision of Ganesh Balakrishnan.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. S. D. Jackson and R. Jain, “Fiber-based sources of coherent mir radiation: key advances and future prospects,” Opt. Express 28(21), 30964–31019 (2020). [CrossRef]
2. B. Srinivasan, E. Poppe, J. Tafoya, et al., “High-power (400 mW) diode-pumped 2.7 μm Er: ZBLAN fibre lasers using enhanced Er-Er cross-relaxation processes,” Electron. Lett. 35(16), 1338–1340 (1999). [CrossRef]
3. S. D. Jackson, T. A. King, and M. Pollnau, “Diode-pumped 1.7-W erbium 3-μm fiber laser,” Opt. Lett. 24(16), 1133–1135 (1999). [CrossRef]
4. T. Sanamyan, “Efficient cryogenic mid-IR and eye-safe Er:YAG laser,” J. Opt. Soc. Am. B 33(11), D1–D6 (2016). [CrossRef]
5. B. Topper, A. Neumann, A. Mafi, et al., “Nonlinear optical properties of lanthanum titanate glasses prepared by levitation melting,” Appl. Phys. Lett. 124(9), 1 (2024). [CrossRef]
6. B. Topper, S. K. Wilke, M. Pettes, et al., “Mid-infrared luminescence properties of erbium and dysprosium doped lanthanum titanate glasses,” Opt. Mater. Express 13(10), 2857–2868 (2023). [CrossRef]
7. B. Topper, A. Neumann, S. K. Wilke, et al., “Site-selective fluorescence and spectroscopic properties of Yb-doped lanthanum titanate glasses,” Int. J. Appl. Glass Sci. (2024).
8. S. K. Wilke, O. L. Alderman, C. J. Benmore, et al., “Octahedral oxide glass network in ambient pressure neodymium titanate,” Sci. Rep. 12(1), 8258 (2022). [CrossRef]
9. S. K. Wilke, A. Al-Rubkhi, C. Koyama, et al., “Microgravity effects on nonequilibrium melt processing of neodymium titanate: thermophysical properties, atomic structure, glass formation and crystallization,” npj Microgravity 10(1), 26 (2024). [CrossRef]
10. J. Allain, M. Monerie, and H. Poignant, “Energy transfer in Er3+/Pr3+-doped fluoride glass fibres and application to lasing at 2.7 μm,” Electron. Lett. 27(5), 445–447 (1991). [CrossRef]
11. D. Knowles, A. Cassanho, and H. Jenssen, “Comparison of Er-and Er, Pr-doped YLF 2.8-μm lasers,” in OSA Annual Meeting, (Optica Publishing Group, 1990), p. ThJJ3.
12. L. Wetenkamp, G. West, and H. Többen, “Co-doping effects in erbium3+-and holmium3+-doped ZBLAN glasses,” J. Non-Cryst. Solids 140, 25–30 (1992). [CrossRef]
13. H. Lin, D. Chen, Y. Yu, et al., “Enhanced mid-infrared emissions of Er3+ at 2.7 μm via Nd3+ sensitization in chalcohalide glass,” Opt. Lett. 36(10), 1815–1817 (2011). [CrossRef]
14. X. Li, P. Zhang, H. Yin, et al., “Sensitization and deactivation effects of Nd3+ on the Er3+: 2.7 μm emission in PbF2 crystal,” Opt. Mater. Express 9(4), 1698–1708 (2019). [CrossRef]
15. Y. Wang, J. Li, Z. Zhu, et al., “Dual function of Nd3+ in Nd, Er:LuYSGG crystal for LD pumped 3.0 μm mid-infrared laser,” Opt. Express 23(14), 18554–18562 (2015). [CrossRef]
16. H. Xia, J. Feng, Y. Wang, et al., “The effects of Ho3+ and Pr3+ ions on the spectroscopic properties of Er3+ doped SrGdGa3O7 crystals used in mid-infrared lasers,” J. Phys. D: Appl. Phys. 48(43), 435106 (2015). [CrossRef]
17. S. Lv, Z. Zhu, Y. Wang, et al., “Spectroscopic investigations of Ho3+/Er3+: CaYAlO4 and Eu3+/Er3+: CaYAlO4 crystals for 2.7 μm emission,” J. Lumin. 144, 117–121 (2013). [CrossRef]
18. J. R. Weber, “The containerless synthesis of glass,” Int. J. Appl. Glass Sci. 1(3), 248–256 (2010). [CrossRef]
19. O. L. Alderman, C. J. Benmore, A. Tamalonis, et al., “Rare-earth titanate melt structure and glass formation,” Int. J. Appl. Glass Sci. 10(4), 463–478 (2019). [CrossRef]
20. A. Schwuchow, S. Unger, S. Jetschke, et al., “Advanced attenuation and fluorescence measurement methods in the investigation of photodarkening and related properties of ytterbium-doped fibers,” Appl. Opt. 53(7), 1466–1473 (2014). [CrossRef]
21. D. Sumida and T. Fan, “Effect of radiation trapping on fluorescence lifetime and emission cross section measurements in solid-state laser media,” Opt. Lett. 19(17), 1343–1345 (1994). [CrossRef]
22. S. Kang, D. He, X. Wang, et al., “Effects of CaO/Al2O3 ratio on structure and spectroscopic properties of Nd3+-doped CaO-Al2O3-BaO aluminate glass,” J. Non-Cryst. Solids 468, 34–40 (2017). [CrossRef]
23. Y. V. Baklanova, O. A. Lipina, A. N. Enyashin, et al., “Nd3+, Ho3+-Codoped apatite-related NaLa9(GeO4)6O2 phosphors for the near-and middle-infrared region,” Dalton Trans. 47(39), 14041–14051 (2018). [CrossRef]
