Introduction

Two micron eye-safe laser sources based on Holmium-doped materials have proven important for a variety of applications ranging from the medical such as stone retropulsion [1] and ablation of hard tissues [2], to remote sensing of atmospheric CO₂ and H₂O using LIDAR [3,4] techniques as well as frequency conversion by optical parametric oscillators [5]. Because of their relatively low gain cross section, early use of Ho³⁺ doped materials during the years of flashlamp pumping required that the gain medium be operated at 77 K (Ho silicate glass [6], Ho³⁺:CaWO₄ [7] and Ho³⁺:YAG [8]). The later development of bright and spectrally-concentrated pump (diode) near infrared sources along with co-doping Ho with Tm permitted room temperature laser operation (e.g., [9]) through absorption at 785nm. Unfortunately, the conversion of the 785 nm pump source through the “two-for-one” relaxation process results in a large waste heat deposition (>25% quantum defect), optical distortion and mechanical stresses in the gain medium thereby unduly limiting the power scaling potential of the laser.

In order to minimize the quantum defect, resonant pumping of the ⁵I₇ level followed by the ⁵I₇→⁵I₈ laser transition is highly desirable, analogous to the low quantum defect resonant pumping of Er³⁺ utilizing the ⁴I_15/2 → ⁴I_13/2 transitions [10]. To date, a number of pump options have been shown to be effective for the resonant pumping of Ho³⁺ in different single crystalline hosts, for example Tm-doped fiber [11] and bulk solid-state lasers [12,13] and semiconductor AlGaIn/AsSb diode lasers [14]. While resonant pumping has been shown to be successful in reducing gain medium temperature excursions in Ho³⁺:YAG, further reduction in temperature gradients can be expected by replacing YAG with a laser host of greater thermal conductivity. One such candidate host is Y₂O₃ (yttria) with a thermal conductivity superior to YAG over the temperature range from ~100 - 300 K [15]. Unfortunately, single-crystalline yttria is difficult to obtain by traditional high temperature melt growth methods. Due to the limited availability of doped yttria materials, in particular Ho³⁺-doped Y₂O₃, very little has been reported on the spectroscopy of Ho³⁺:Y₂O₃ [16]. However, recent developments in laser ceramics technology [17] have begun to increase the availability of sesquioxide laser host materials.

This paper presents the spectroscopic data (absorption, stimulated emission and gain cross section spectra) pertinent to resonant pumping of the Ho³⁺ ion based on the ⁵I₇ → ⁵I₈ transitions along with the first demonstration of laser operation of Ho³⁺:Y₂O₃ at 2.12 µm. To the best of our knowledge, this is the first reported laser operation of Ho³⁺-doped Y₂O₃. Due to the marginal quality of currently available Ho³⁺:Y₂O₃ laser material we chose to operate the Ho³⁺:Y₂O₃ laser at cryogenic temperature in our first experiments. The cryo-cooled Ho³⁺:Y₂O₃ laser as resonantly diode-pumped at ~1.93 µm delivered up to 2.5 W of CW output power at ~2.12 µm with a slope efficiency of 35% with respect to absorbed power and a beam propagation factor, M², of ~1.1.

Ho³⁺:Y₂O₃: sample preparation and spectroscopy

The transparent Ho³⁺-doped Y₂O₃ ceramics used in our experiments were prepared by solid-state synthesis followed by cold isostatic compression, sintering and then hot isostatic pressure treatment at 1600 °C. Scanning electron microscope imagery showed the sample to have an average grain size of ~5µm. Spectroscopic measurements were made at 77 K and 300 K. Absorption measurements were made on a Ho³⁺(1%):Y₂O₃ sample with 0.2 cm⁻¹ resolution using a PerkinElmer 2000 FTIR spectrometer. Measurements of the emission spectrum were made on a powdered Ho³⁺(1%):Y₂O₃ sample using a Horiba Fluorolog spectrometer (iHR320) with 0.3 nm spectral resolution and a Judson J23-5I-R02M-2.4 uncooled extended sensitivity InGaAs (out to ~2.4 µm) detector. A powdered form of sample was used in order to avoid effects of re-absorption on the spectra and radiation trapping on upper laser level lifetime measurements. The sample was excited into the ⁵F₃ manifold by a frequency doubled Spectra-Physics Tsunami Ti:Sapphire laser operated at 910 nm. The 455 nm wavelength is one of the best wavelengths to use for powder excitation of Ho³⁺:Y₂O₃ as the absorption cross section in the ⁵F₃ manifold is at least twice that of any other lower lying manifold. Fluorescence decay measurements of the ⁵I₇ upper laser level at 77 K and 300 K showed the decay to be single-exponential with a lifetime of τ = 11.3 ms and 12.3 ms respectively.

The ⁵I₈ →⁵I₇ absorption cross section spectrum of Ho³⁺:Y₂O₃ is presented in Fig. 1 . The cross section was derived from the measured absorption coefficient using the effective dopant ion number density, N_eff = 2.01 × 10²⁰ cm⁻³ per 1% of Ho³⁺ ion concentration. N_eff accounts for the fact that only ~3/4 of the actual dopant ion density occupies Y sites with C₂ symmetry in the Y₂O₃ lattice. The remaining ¼ of the available Y sites, those with C_3i symmetry, give substantially weaker transitions than those from C₂ sites due to inversion symmetry, and thus are expected to contribute relatively little to the spectra [18].

 

Fig. 1 ⁵I₈ → ⁵I₇ absorption cross section of Ho³⁺:Y₂O₃ at 77 K overlaid with diode laser (pump) spectral output. The laser diode spectral output is shown in arbitrary units as recorded at a pump current of 9A and coolant temperature 25 °C.

Download Full Size | PDF

As can be seen in Fig. 1, the absorption spectrum of Ho³⁺:Y₂O₃ shows a grouping of peaks in the range of 1930-1944 nm with a maximum cross section of σ_A = 4.75 × 10⁻²⁰ cm² at 1930.5 nm. The absorption peaks were fully resolved with a FWHM of 1.0 nm. It is of practical interest to note that the four primary absorption peaks around 1935 nm can be spanned by the emission spectrum of a typical non-narrowed indium phosphide based laser diode source as shown in Fig. 1, which, thus, can be effectively utilized for resonant pumping

The stimulated emission cross section spectrum, σ_SE, is shown in Fig. 2 . It was derived by using a combination of reciprocity and Fuchtbauer-Ladenburg methods [19]. Since the reciprocity method works only where absorption is measurable, it was used only in the wavelength range of 1850-1990 nm. Conversely, as the F-L method functions best where re-absorption is minimal, σ_SE was calculated by that means only in the wavelength range of 1950-2150 nm. The ranges were combined with minimal scaling of the FL results as determined by overlapping values in the 1970 nm region in Fig. 2.

 

Fig. 2 ⁵I₇ → ⁵I₈ stimulated emission cross section of Ho³⁺:Y₂O₃ at 77 K.

Download Full Size | PDF

As shown in Fig. 2, in the ground-state absorption-free region has two equally strong emission peaks occur at 2087 nm and 2119 nm with σ_SE(77 K) = 1.0 × 10⁻²⁰ cm². However, the calculated gain cross section, σ_g, indicates that the Ho³⁺:Y₂O₃ laser should favor the 2119 nm wavelength over 2087 nm by a small margin. The gain cross section can be derived from

 

Fig. 3 Calculated gain cross section of Ho³⁺:Y₂O₃ at 77 K.

Download Full Size | PDF

For purposes of comparison to the 77 K spectra, the room temperature (300 K) absorption and stimulated emission spectra are presented in Fig. 4 . It is clear from Fig. 4 that while there is a reduction in the peak absorption cross sections of the principal pump lines around 1.93µm, a broadening of the lines continues to allow for strong absorption of un-narrowed pump diodes. However from the aspect of gain, the stimulated cross section (peak) is reduced by one half to σ_SE(300K) = 4.5 × 10⁻²¹ cm² at 2119 nm. We therefore expect a severe reduction of the gain cross section at 300K, and in fact, calculation of the gain cross section indicates that the population inversion needs to exceed 20% before σ_g exceeds 10⁻²² cm² at room temperature. However for both temperatures, gain calculations show that Ho³⁺:Y₂O₃ will operate at ~2.12 µm.

 

Fig. 4 Absorption and stimulated emission cross sections of Ho^3+:Y₂O₃ at 300 K.

Download Full Size | PDF

Laser setup

The Ho³⁺:Y₂O₃ ceramic laser was CW pumped by a fiber coupled (14 W out of a 400 µm core, NA ~0.2) indium phosphide based laser. The center wavelength and the FWHM of the pump source spectral output were highly pump current dependent (e.g. 1906 nm, 7 nm FWHM at 4W (3A) output and 1934nm, 14 nm FHWM at 14 W (9A) – at a coolant temperature of 25 C). The fiber output was collimated and then focused onto a 3.5 × 3.5 mm², 10 mm long, Ho³⁺(3%):Y₂O₃ rod, which was anti-reflection (AR) coated (<0.1% per side at 2119 nm, < 0.4% at 1930 nm). The sample was conductively cooled through contact of an Indium layer to an oxygen-free copper block. The copper block was cooled by the cold finger of a standard liquid nitrogen boil-off cryostat. The laser cavity was formed by a dichroic (AR at ~1.93 µm, highly reflective at ~2.1 µm), 750 mm radius-of-curvature (ROC) spherical mirror and a selection of 500 mm ROC partially reflective outcouplers (~99%, 95%, 90% reflectivity at the laser wavelength of 2119 nm). The laser cavity was 250 mm long and the cavity mode was calculated to be ~800 µm (1/e²) diameter at ~2 µm. The 1.93 µm pump beam was measured to have a diameter at focus of 550 µm and was projected not to exceed 800 µm diameter anywhere inside the 10 mm long rod. Measurement of the pump profile was made by re-imaging the focused beams onto a 124 × 124, 100 µm pitch pyroelectric array (Spiricon Pyrocam III) camera. The fact that the focused pump beam remained within the laser cavity mode assured an optimal utilization of absorbed pump power.

Experimental results and discussion

In order to measure the Ho³⁺:Y₂O₃ laser output vs. input characteristic, the fiber coupled diode laser pump module power was varied by changing the diode laser driver current. Because the pump spectral output varied greatly with drive current (~2.8 nm/W), the fraction of light absorbed by the Ho³⁺(3%):Y₂O₃ sample was measured for all diode drive currents. For the range of drive currents where lasing occurred, the absorption was found to vary progressively from 16% at lowest (4 W) pump power to 72% at highest (14 W). The fraction of absorbed pump power increased with current as the pump diode emission spectrum shifted to the 1930 nm absorption peaks. We estimate that the fraction of absorbed power under non-lasing conditions is a good approximation for the fraction under lasing condition as Ho:Y₂O₃ is a four level system at ~80 K. The Ho³⁺(3%):Y₂O₃ laser output at ~2.12 µm is plotted in Fig. 5 as a function of absorbed pump power. The results indicate that the maximum slope efficiency of ~35% with respect to absorbed power occurs with the 10% output coupler. While higher slope efficiency can be expected considering the ~90% quantum defect limit based on pumping and emission wavelengths, the optical losses of the underdeveloped Ho³⁺:Y₂O₃ laser ceramic lead to greatly reduced laser performance.

 

Fig. 5 Resonantly diode pumped laser performance of Ho³⁺:Y₂O₃ at 77K.

Download Full Size | PDF

Two efforts were made to quantify the scattering loss within the crystal, which limited achieved laser efficiency. The first used a 1.52 µm HeNe laser beam to probe the 3.5 mm (uncoated) width of the crystal. It was determined that an average optical loss of α = 0.07 ± 0.02 cm⁻¹ was present in the Ho⁺³(3%):Y₂O₃ ceramic which creates a loss of 7% over the 10 mm crystal length. By adding the measured 1.5% single-pass loss of the cryostat windows at 2119 nm to the crystal optical losses, we determined that the laser cavity experienced close to ~8.5% single-pass cavity loss. A second effort was made to deduce the single pass cavity loss through one dimensional (1-D) laser modeling [21]. Adjustment of the optical loss in the laser model to the observed slope efficiencies allowed us to deduce a similar overall cavity single pass loss of 9-11%.

As beam quality is of interest, the M² value of the 2.12 µm laser was measured under 1.75 W, CW operation. As reproduced in Fig. 6 , the beam quality was observed to have an M² value of ~1.1 in the orthogonal radial directions of x and y. The Spiricon Pyrocam III beam profiler was used to record the laser emission intensity profile. The beam was focused by a 125 mm focal length lens and re-imaged onto the pyrocam camera resulting in a pixel resolution of 25 µm. An example of the beam profile at focus is shown as an inset in Fig. 6.

 

Fig. 6 Beam quality (M_x ² = M_y ² ~1.1) measurement of the 2.12 µm laser under CW, 1.75 W operation. An example of the beam profile at focus has been included as an inset.

Download Full Size | PDF

The laser emission spectra at two output powers of 1.75 W and 2.1 W were measured by a Yokogawa AQ6375 optical spectrum analyzer with a spectral resolution of 0.05 nm as presented in Fig. 7 . It is evident from Fig. 7 that the laser operated with two distinct emission profiles around a central wavelength of about 2118.8 nm. Both spectra exhibit longitudinal modes of spectral width, δλ_FWHM = 0.15 nm, and spacing of Δλ = 0.47 nm. The longitudinal mode spacing results from internal reflection of the 3 mm thick cryostat windows, while the smaller Fabry-Perot modes also visible on the spectra are due to the cryostat window separation of 60 mm. In both cases(1.75 W and 2.1 W), the laser emission span no more than 1.5 nm which is less than the gain emission cross section width of ~4 nm as shown in Fig. 3.

 

Fig. 7 Ho³⁺:Y₂O₃ laser emission spectrum at 77K with evidence of cryostat window induced Fabry-Perot modes.

Download Full Size | PDF

Conclusions

Reported here is what we believe to be the first laser operation based on Ho³⁺-doped Y₂O₃. Resonantly diode-pumped (at ~1.93 µm) laser operation of Ho³⁺:Y₂O₃ was demonstrated at 2.119 µm with a slope efficiency of 35% with respect to absorbed power, and 2.5 W of CW output at 77K and with a beam propagation factor of M² ~1.1. The absorption, emission and gain cross sections of Ho³⁺:Y₂O₃ and the lifetime of the ⁵I₇ - ⁵I₈ transition are presented for 77 and 300K with measured fluorescence lifetimes of 11.3 ms and 12.3 ms respectively. The scattering loss of the Ho(3%):Y₂O₃ ceramic has been measured to be ~0.07cm⁻¹ and the overall intracavity loss figure estimated as approximately 10% per pass. Considering the substantial losses present in these first experiments we expect to achieve a greater laser efficiency (>70%) once the single pass cavity loss is reduced to a reasonable 1% value.