Study on Spectral Properties and Mid-Infrared Laser Performance of Er, La:CaF₂ Crystals

Abstract

Er³⁺-doped fluorite crystals, including CaF₂ and SrF₂, are considered as attractive laser gain materials in the mid-infrared (MIR) region with merits of high laser efficiency as well as low doping concentration. In this work, a series of Er, La:CaF₂ crystals were grown and the modulation effect of co-doping La³⁺ ions on the spectral properties and mid-infrared laser performance was investigated. It was found that introducing La³⁺ ions can effectively manipulate the coordination environment of Er³⁺ ions embedded in CaF₂ crystal, thus modulating the shape and intensity of absorption and emission bands. On the other hand, La³⁺ ions can partially substitute Er³⁺ sites in the clusters to form mixed clusters, which affects the energy transfer processes between Er³⁺ ions as well as ~3 μm laser performance, which is dominated by energy transfer up-conversion (ETU) processes between Er³⁺ ions. By co-doping La³⁺ ions into Er:CaF₂ crystal at an appropriate concentration, the spectral parameter modulation can be achieved while maintaining a high MIR laser efficiency.

1. Introduction

A near 3 μm laser based on laser diode (LD)-pumped Er-doped gain materials has attracted increasing attention for significant applications in cosmetic medicine, soft tissue ablation, and gas sensing [1,2,3]. More importantly, it can be used directly as a pump source for Fe²⁺:ZnSe to achieve lasing at a wavelength range of 4~5 μm [4], which is critical for the advanced infrared countermeasure system. Compared to optical parameter oscillation (OPO) or optical parameter amplification (OPA) schemes, the Er laser-pumped Fe²⁺:ZnSe system possesses higher integration and improved robustness, and more importantly its output power is able to scale to the hundred-watts level with Master Oscillator Power-Amplifier (MOPA) technology [5]. Thus, there have been consistent and increasing efforts to realize a high-power or high-energy Er-based near 3 μm laser in the past decades.

It is well known that near 3 μm laser oscillation derived from Er³⁺: ⁴I_11/2 → ⁴I_13/2 transition is self-terminated because of the longer lifetime of the upper energy level ⁴I_13/2 than that of the lower energy level ⁴I_11/2 [6]. To overcome this unfavorable effect, high concentration doping is adopted for Er-doped gain materials in order to take advantage of the ETU1 process (⁴I_13/2 + ⁴I_13/2 → ⁴I_9/2 + ⁴I_15/2) which can accelerate depopulation of ⁴I_13/2, and meanwhile recycle the excitation energy of ⁴I_13/2 to multiply the population of ⁴I_11/2 [7,8], as illustrated in Figure 1a. On the other hand, introduction of numerous dopants will degrade the thermal conductance of laser gain crystals [9,10] and result in severe thermal lens effect due to the intense absorption of the pump light, which hinders the power scaling of Er-doped crystals [11,12].

Er³⁺-doped fluorite crystals, including CaF₂ and SrF₂, stand out for the unique self-aggregation behavior or clustering effect [13] of the trivalent rare earth (RE³⁺) ions embedded in the crystals (shown in Figure 1b), which enables this kind of gain crystals to achieve high laser efficiency at a rather lower doping level. Benefiting from the ionic clustering-induced enhancement of the ETU1 process, Fan et al. and Liu et al. realized a highly efficient ~2.8 μm continuous-wave (CW) laser with a slope efficiency above 40% by 976 nm LD end-pumped Er:SrF₂ and Er:CaSrF₂ crystals [14,15]; this efficiency surpasses the Stokes limitation and still keeps the highest record for LD-pumped Er laser in the 3 μm wavelength range. In addition, a CW MIR laser with hundreds of milliwatts output power was also obtained in extremely light-doping (0.3 at.%) Er:CaF₂ crystals [16], which breaks the long-standing consideration that a ~2.8 μm CW laser is unable to run in low concentration Er-doped materials. Based on the merit of low doping concentration, a 5.04 W CW laser at 2799 nm was obtained in a 2 at.% Er:CaF₂ crystal [17]. The first semi-conductor saturable absorber mirror (SESAM) mode-locked laser operation of Er-doped bulk materials was realized in Er:CaF₂-SrF₂ mixed crystal [18].

The ~3 μm laser performance of Er-doped crystals is dominated by two different energy transfer up-conversion processes, that is, ETU1: ⁴I_13/2 + ⁴I_13/2 → ⁴I_9/2 + ⁴I_15/2 and ETU2: ⁴I_11/2 + ⁴I_11/2 → ⁴F_7/2 + ⁴I_15/2; the former is desired and the latter is adverse to the laser oscillation. These two processes both occur with lasing and their competitive strength varies with the Er³⁺ coordination environment [19]. The strength regulation of these two ETU processes is generally realized by changing the doping concentration of Er³⁺ ions in Er:LiYF₄ [20], Er:YSGG [21], Er:BaY₂F₈ [22], and so on. In contrast, for Er³⁺-doped fluorite crystals, introducing inactive rare earth ions to manipulate the cluster structure should be another more effective way to achieve this regulation. In fact, co-doping optically inert RE³⁺ ions (such as Y³⁺, La³⁺, Gd³⁺, Lu³⁺ etc.) has been demonstrated to modulate the site structure and spectral properties of Nd³⁺ ions embedded in CaF₂, SrF₂, and their mixed crystals [23,24,25]. However, most works are focused on an Er³⁺ singly doped CaF₂/SrF₂ system at present, and the effect of co-doping site-modulated ions remains unknown.

In this work, a series of Er³⁺, La³⁺ co-doped CaF₂ crystals were grown to investigate the modulation effect of co-doping La³⁺ ions on the spectral properties and MIR laser performance. It was found that introduction of La³⁺ ions into Er:CaF₂ crystal could change the site structure of Er³⁺ ions within the ionic clusters, reshaping the absorption/emission band and enhancing the emission cross-section and lifetimes of ~3 μm laser energy levels. In particular, a moderate La³⁺ concentration such as 2 at.% can make improvements that do not impede ETU processes, resulting in a low laser threshold and high laser efficiency.

2. Materials and Methods

The 5 at.% Er, x at.% La:CaF₂ (x = 0, 2, 5, 10) single crystals were successfully grown by temperature gradient method. Crystalline ErF₃, LaF₃, and CaF₂ with a purity of 99.99% were ground and well mixed according to a stoichiometric ratio. An additional 1 wt.% PbF₂ powder was added to serve as oxygen scavenger. The mixed powder was then loaded into a muti-hole graphite crucible with the caps dilled a Φ1 mm hole in the center. At first, the vacuum in the sealed furnace chamber was pumped to be lower than 1 × 10⁻³ Pa. Then, the crucible was heated to 200 °C at the rate of 30 °C per hour and the temperature was kept constant for 5 h to get rid of the absorbed water of the raw powder. The crucible was then heated to 800 °C and kept for 12 h to eliminate the oxygen in the chamber through the reaction of PbF₂ + O₂ → PbO + F₂. The crucible was further heated to 1420 °C and maintained for 15 h to ensure the homogeneity of the melt. After that, the temperature declined at a cooling rate of 0.8 °C/h during the period of crystallization, and finally the crucible was cooled down to room temperature at a rate of 15 °C/h. Crystal chips with a thickness of 1 mm and blocks with a cross-section of 3 × 3 mm² and a length of 10 mm were cut from the individual as-grown boules, then the end faces of the plates and blocks were finely polished. The thin chips were prepared for spectral measurements and the crystal blocks were used as gain elements in laser experiments.

The powder X-ray diffraction (PXRD) patterns were collected by a D8 ADVANCE high-resolution X-ray diffractometer (Bruker, Karlsruhe, Germany). The real doping concentrations of Er³⁺ and La³⁺ ions were determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES, Agilent 725). Absorption spectra were measured using a Cary 5000 UV/VIS/IR spectrophotometer (Agilent, Santa Clara, CA, USA) with an integration time of 0.2 s and a step of 0.1 nm. Emission spectra and fluorescence decay measurements were carried out by an FLS 1000 fluorescence spectrometer (Edinburgh Instruments, Livingston, UK) where the excitation sources included a steady-state xenon lamp, a microsecond flash lamp, and a 980 nm semiconductor laser which can work in CW and pulsed mode. The detectors used involved visible and infrared photomultiplier tubes (PMTs) and a liquid nitrogen cooled InSb infrared photodiode. All measurements were carried out at room temperature.

The laser experiments were carried out with a linear plano-concave resonator as depicted in Figure 2. A fiber-coupled 976 nm LD was used as pump source. The core diameter of the fiber was 105 μm with a numerical aperture (NA) = 0.22. The pump beam was collimated and focused by a couple of convex lenses with the expansion ratio of 1:2. The input mirror (IM) was a concave lens with curvature radius of −50 mm and coated with high transmittance (HT) for the pump beam and high reflectivity (HR) at 2.7~2.95 μm. A series of plane lenses with different transmittance of 1%, 2%, and 3% for laser beam were utilized as the output mirror (OM). The laser beam was reflected to the power meter by a dichroic mirror with HR for the laser beam and HT for the pump beam so the laser output power could be measured.

3. Results and Discussion

Figure 3a shows the as-grown 5 at.% Er, 2 at.% La:CaF₂ crystal ingot which was transparent and without any cracks or inclusions. For fluorite crystals, the cellular structure is apt to occur especially in the tail of the ingots; the chips and blocks shown in Figure 3c, d were selected from the homogeneous parts of the as-grown boules which are lacking the cellular structure. The actual concentrations of dopants were obtained by ICP-AES tests and the results are summarized in

Table 1. Summary of the real doping concentrations of the crystals.

Crystals	Er3+		La3+
	at.%	1021 ions/cm3	at.%	1021 ions/cm3
5 at.% Er:CaF2	4.74	1.16	/	/
5 at.% Er, 2 at.% La:CaF2	4.67	1.11	2.12	0.449
5 at.% Er, 5 at.% La:CaF2	4.82	1.13	6.22	1.38
5 at.% Er, 10 at.% La:CaF2	5.25	1.21	11.41	2.66

. There are slight deviations between the nominal and actual doping concentrations because of segregation during crystal growth. To check the phase purity of the crystals, PXRD measurements were performed and the results are presented in Figure 4a. All the diffraction peaks belong to the CaF₂ phase, which demonstrates that the fluorite structure can be retained with even more than 15 at.% RE³⁺ dopants introduced. However, it was found that the series peaks shift to low angle with an increase in La³⁺ dopants, indicating that lattice expansion occurs when La³⁺ (1.30 Å) ions substitute original sites occupied by Ca²⁺ (1.26 Å) ions as well as when the equivalent amount of F⁻ ions enters the interstitial site. Many calculations and experimental results reported indicate that the introduced RE³⁺ ions and charge-compensated interstitial F⁻ in CaF₂ will congregate spontaneously and form locally disordered ionic clusters under the combined effect of Coulomb force and lattice stress [26,27]. In Figure 4b, it can be seen that the variation of lattice constant on La concentration failed to follow Vegard’s law well, maybe due to the inconstant Er doping concentration.

The spectral characteristics were systematically investigated to analyze the modulation effect of co-doping La³⁺ ions. Figure 5 shows the absorption cross-sections (σ_abs) corresponding to Er³⁺: ⁴I_15/2 → ⁴I_11/3 and ⁴I_13/2 transitions. The former transition can be directly excited by commercial ~980 nm LDs, thus it is paid more attention. For the singly doped crystal it can be seen that this absorption band consists of two peaks centered at 967.3 nm and 981.2 nm, and each peak has a shoulder which emerges at 970 nm and 979 nm, respectively. With La³⁺ dopants increasing, the 967.3 nm peak shows slight red-shift and the σ_abs is significantly reduced by 40%, from 2.28 × 10⁻²¹ cm² to 1.44 × 10⁻²¹ cm², which reduces to be lower than that of the shoulder for 10 at.% La³⁺ co-doped crystal. On the other hand, for the peak at 981.3 nm, a more obvious blue-shift occurs with La³⁺ concentration; the σ_abs shows a small increase at 2 at.% La³⁺ concentration, then decreases by about 20% (from 2.09 × 10⁻²¹ cm² to 1.70 × 10⁻²¹ cm²), and the original shoulder becomes the new peak. Finally, the two peaks converge into one with the peak center at 975.9 nm. There are two factors that can account for the absorption band evolution: One is that the fraction of the active centers corresponding to the two original absorption peaks declines with La³⁺ concentration; in addition, La³⁺ and Er³⁺ ions doped in CaF₂ crystal will form mixed clusters where the strength of crystal field applied to Er³⁺ ions becomes weaker than that in pure Er³⁺ clusters, leading to a reduction of Stark level splitting of Er³⁺: ⁴I_11/2. The similar change also occurs to the absorption band at near 1550 nm: the σ_abs first slightly increases at 2 at.% La³⁺ concentration and then declines, accompanied with the gradual convergence of the original structured band.

The emission cross-sections (σ_em) near 2.8 μm were calculated according to the Judd–Ofelt theory and the Füchtbauer-Ladenburg formula [28], and are illustrated in Figure 6. The σ_em at 2729 nm also shows a rise for 2 at.% La³⁺ co-doped sample (from 7.3 × 10⁻²¹ cm² to 7.8 × 10⁻²¹ cm²), then a decline with La³⁺ concentration increasing. The broad emission band in the long wavelength range overall decreased, which results in the width of this emission band narrowing.

We further measured the lifetimes of Er³⁺: ⁴I_11/2 and ⁴I_13/2 levels. The thin samples were excited by a 980 nm semiconductor laser and the pulse width was about 50 μs. To prevent the influence of self-absorption on the intrinsic fluorescence decay, the samples for lifetime measurements were processed into the thin chips with a thickness of 1 mm, and the signals were monitored at 1002 nm and 1570 nm, which is away from the corresponding absorption peaks. The decay curves of Er³⁺: ⁴I_11/2 level are summarized in Figure 7a; they all can be well-fitted with a single exponential function, and the fitted lifetimes are 6.7 ms, 7.5 ms, 7.8 ms, and 7.9 ms, with increase of La³⁺ concentration. For the decay curves of ⁴I_13/2, because of the population compensation by Er³⁺: ⁴I_11/2 → ⁴I_13/2 transition under 980 nm excitation, there is an obvious rising edge in the first 10 ms of decay curves, as shown in Figure 7b. These curves were fitted by the formula put forward in Ref [29] and the lifetimes were calculated to be 11.1 ms, 14.3 ms, 14.5 ms, and 15.3 ms for 0 at.%, 2 at.%, 5 at.%, and 10 at.% La³⁺-doped samples. The lifetimes were prolonged with 2 at.% La³⁺ introduced, and the rise in lifetime becomes smaller when further increasing La³⁺ concentration.

As mentioned before, introducing La³⁺ ions changes the composite of the original RE³⁺ clusters, which influences the coordinate structure of Er³⁺ ions residing in the clusters; on the other hand, the pure Er³⁺-clusters could be partially “broken up”, leading to the cooperative up-conversion between Er³⁺ ions being weakened. In order to investigate the “weakening effect” of La³⁺ dopants, the decay curves of ⁴S_3/2 were measured. It has been demonstrated that the fluorescence originated from the ⁴S_3/2 level is sensitive to Er³⁺ concentration and easy to quench by Er³⁺: ⁴S_3/2 + ⁴I_15/2 → ⁴I_11/2 + ⁴I_13/2 process [30], thus the lifetime of ⁴S_3/2 can reflect the interaction strength between Er³⁺ ions. The results are shown in Figure 7c; all the decay curves fit well with the single exponential function and the fitted lifetimes were 16.8 μs, 17.4 μs, 21.4 μs, and 26.5 μs, as summarized in Figure 7d. With the La³⁺ concentration increasing, the green fluorescence of ⁴S_3/2 with a radiation lifetime of several hundred microseconds [31] is seriously quenched for all crystal samples and the rise of the lifetime is inconspicuous with La³⁺ concentration. It is worth noting that for the 2 at.% La³⁺-doped sample, the lifetime increase is negligible, which is not consistent with that at ~1 μm or ~1.5 μm. This indicates that introducing La³⁺ into 5 at.% Er:CaF₂ crystal only has a moderate influence on the interionic energy transfer processes; the spectral modulation, including reshaping the absorption/emission band and increasing emission cross-sections and lifetimes of laser energy levels, is mainly attributed to the change of the site structures surrounding the embedded Er³⁺ ions. In fact, it has been demonstrated that La³⁺ ions tend to aggregate with cubic coordination in CaF₂ crystal, whereas Er³⁺ ions prefer to form clusters with tetragonal antiprism sublattice [32]. Therefore, when La³⁺ ions partially substitute Er³⁺ sites within clusters to form mixed ones, the extra lattice distortion with the structural characteristics of La³⁺ clusters will occur and achieve the local structure manipulation.

We also tested the laser performance of the Er³⁺, La³⁺ double doping crystals based on the setup previously mentioned, and the results are presented in Figure 8. For 5 at.% Er, 2 at.% La:CaF₂ crystal, higher laser efficiencies were obtained when T = 2% and 3% OMs were utilized. The highest slope efficiency was 36.6% and the maximum output power was 0.92 W with a T = 3% OM. When using a 2% T OM, the slope efficiency was 35.6% with an output power of 0.9 W. The laser thresholds, increasing with the transmittance of the OM, were 31 mW, 98 mW, and 153 mW for the OM with T = 1%, 2%, and 3%. When La³⁺ concentration increased to 5 at.%, neither the slope efficiency or the maximum output power degraded; the highest slope efficiency was 29.4% with an output power of 0.435 W for T = 3% OM, and the laser threshold became higher, 146 mW for T = 1% OM. The degraded laser performance is partially due to the weakened ETU process, and the lower thermal conductivity with a higher doping concentration, resulting in more severe thermal effect that also impedes efficient laser operation. For the 5 at.% Er, 10 at.% La:CaF₂ crystal, it is difficult to stabilize the laser oscillation because of the serious thermal effect. We also failed to perform the laser experiment on 5 at.% Er:CaF₂ crystal due to the undesired crystal quality; however, the laser performance of 5 at.% Er:CaF₂ crystal reported in Ref. [33] can serve as a reference. As shown in

Table 2. Summary of ~3 μm CW laser performance of LD-pumped Er-doped crystals.

Crystals	Er concentration (at.%)	Slope Efficiency (%)	Output Power (W)	Year
Er:LiYF4	15	35	0.37	1996 [20]
Er:BaY2F8	10	32	0.16	1997 [34]
Er:Lu2O3	7	36/27	1.4/5.9	2012 [35]
Er:YGG	10	35.4	1.38	2018 [36]
Er:SrF2	3	41	1.06	2018 [14]
Er:YSGG	28.3	31.5	0.5	2019 [37]
Er:YAP	5	30.6	6.9	2020 [38]
Er:CaF2	5	37.9	0.7	2022 [33]
Er,La:CaF2	5	36.6	0.92	This Work

, the laser power and slope efficiency were comparable for 5 at.% Er, 2 at.% La:CaF₂ and 5 at.% Er,:CaF₂, which demonstrated that introduction of moderate La³⁺ dopants would not degrade laser performance for 5 at.% Er,:CaF₂.

4. Conclusions

This work demonstrates a method to modulate the spectral parameters concerning ~3 μm MIR laser performance of Er³⁺-doped fluorite crystals. With La³⁺ concentration increasing, the absorption and emission bands are reshaped and the structured bands converge; meanwhile, the emission cross-sections and laser energy lifetimes are enhanced. By analysis of the fluorescence decay behavior of the quenching energy level ⁴S_3/2, it is revealed that co-doping La³⁺ ions would not have a significant influence on the energy transfer process between Er³⁺ ions, and manipulation of the site structure of RE³⁺ ionic clusters plays the dominant role. An efficient CW MIR laser with a slope efficiency of 36.6% and an output power of 0.9 W is achieved in 5 at.% Er, 2 at.% La:CaF₂ crystal. Further increasing La³⁺ concentration would degrade the laser performance due to a weakened ETU process and reduced thermal conductivity. Therefore, the spectral parameters of Er-doped fluorite crystals can be modulated through optimizing the type or concentration of the co-doped optically inert RE³⁺ ions.