Yean Sheng Yong is a PhD student in the MESA+ research group Optical Sciences. His supervisor is Jennifer Herek.
potassium double tungstate waveguides with high ytterbium concentration for optical amplification
This thesis explains the research work concerning high Yb3+ concentration KRE(WO4)2 waveguide layers, where RE stands for Gd, Lu, Y, or the combination of these elements, which are catered for optical amplification purpose. Prior to this work, KRE(WO4)2 high-gain waveguide amplifiers based on epitaxial layer with 47.5 at.% Yb3+ had been achieved. In-depth investigations on the optical properties of the epitaxial layers with > 50 at.% Yb3+ have been conducted in this work via experimental and numerical means to evaluate the potential of these layers for the realization of waveguide amplifiers with superior net gain per unit length performance.
As the luminescent lifetime of high Yb3+ concentration material is typically hampered by radiation trapping effect which elongates the measured lifetime, a novel confocal lifetime measurement setup has been devised and tested. The setup is based on the concept of confocal volume detection which requires i) a pump volume which is as small as possible, and ii) effective discrimination of luminescence originated out of the non-confocal region from the detection system. As compared to other existing approaches being used to mitigate radiation trapping effect, such as measuring on diluted powdered sample or measuring with pinholes followed by an extrapolation procedure, the proposed setup allows direct measurement of lifetime in a non-destructive manner. The setup has been used to determine the lifetime on KRE(WO4)2:Yb3+ epitaxial layers with Yb3+ ranging from 1.2 at.% to 76 at.%. The measured lifetime results fall within the range of reported lifetime values measured using other approaches. Besides, effective suppression of radiation trapping effect using the proposed setup has been demonstrated. The lifetime obtained from sample with 1.2 at.% is 245 ± 3 µs, whereas the lifetime values for KGd0.43Yb0.57(WO4)2 and KLu0.24Yb0.76(WO4)2 are 228 ± 10 µs and 222 ± 9 µs, respectively. Hence, concentration dependent lifetime quenching is rather weak even though these layers exhibit high amount of Yb3+ ions. Apart from that, power dependence of the luminescence decay has been studied. It is found that the time-resolved luminescence curves exhibit non-exponential decay, which indicates the presence of energy transfer upconversion (ETU) process.
Absorption measurements have been performed on KGd0.43Yb0.57(WO4)2 and KLu0.24Yb0.76(WO4)2 to determine the transition cross-sections in these samples. The impact of polarization disorientation on the measured peak absorption has been quantified. Besides, special care has been taken to reduce stray light in order to resolve the absorption peak of KLu0.24Yb0.76(WO4)2 with high absorbance. The cross-sections determined from KGd0.43Yb0.57(WO4)2 and KLu0.24Yb0.76(WO4)2 are found to be similar to those of reported bulk KRE(WO4)2:Yb3+ crystals. The maximum absorption cross-section and emission cross-sections in these films are about 1.3 × 10-19 cm2 and 1.6 × 10-19 cm2, respectively. The difference of cross-section spectra obtained from KGd0.43Yb0.57(WO4)2 and KLu0.24Yb0.76(WO4)2 is marginal even though the Yb3+ concentration in these layers differs by 19 at.%.
Temperature-dependent measurements have been performed on KGd0.43Yb0.57(WO4)2 to quantify the change of spectroscopic parameters with temperature. This is because intensely pumped Yb3+-doped devices often operate at elevated temperature primarily due to the conversion of energy difference between the pump and luminescence photons into heat. It has been confirmed that the lifetime has no significant temperature dependence. On the other hand, a strong dependence of the absorption cross-section, especially on the major absorption peaks at ~932 nm and ~981 nm, on sample’s temperature has been detected. By increasing the temperature on the sample from 20 ˚C to 170 ˚C, a reduction of cross-section by ~40% and ~52% has been determined for the absorption peak at ~932 nm and ~981 nm, respectively. With the aid of a simple model, the reduction of peak absorption cross-section with increasing temperature can be explained by two effects, the reduced fractional population of the relevant Stark level and the linewidth broadening. When the temperature is increased from 20 °C to 170 °C, the fractional population of the relevant Stark level is reduced by ~18%. The linewidths of the decomposed absorption peaks near 932 nm and 981 nm is ~1.37 and ~1.72 times broader at 170 °C than at 20 °C, respectively.
Owing to the strong dependency of transition cross-sections to device’s temperature, a numerical gain model taking into account the pump-induced thermal effects has been established. In addition, pump-probe measurements have been conducted to examine the pump absorption and signal gain in KGd0.43Yb0.57(WO4)2 epitaxy layer. The comparison of the experimental and numerical pump absorption results reveals non-saturation pump absorption behavior. Such effect has been reported in other high active ion concentration materials and it is attributed to separate class of ions which is rapidly quenched from the excited state. By assuming 17% of Yb3+ ions are quenched in addition to ETU process, the calculated result from the numerical model match reasonably to the experimental data. Apart from the pump absorption, the temperature profiles of the sample have been modeled as well. The results show that at the highest launched pump power applied in the gain experiment, the average temperature within the pumped volume is ~65.5 ˚C. The maximum net gain achieved in KGd0.43Yb0.57(WO4)2 is calculated as 2.62 dB, or 817 dB/cm. Two gain limiting factors, which are the population inversion and the temperature of the device, have been identified by comparing the effective gain cross-sections achieved in the sample to theoretical gain cross-section versus temperature curves. Assuming that the temperature of the sample can be reduced to 30 ˚C via active cooling, the net gain per unit lengths could be enhanced by ~30%.
In overall, it has been demonstrated experimentally that KRE(WO4)2:Yb3+ epitaxial layers with Yb3+ > 50% exhibit spectroscopic properties which are very similar to those of bulk KRE(WO4)2:Yb3+ crystals. This indicates that these KRE(WO4)2:Yb3+ layers can be utilized for amplifiers and lasers. Nevertheless, the ETU process, quenched ions, and thermal effects would affect the gain achieved in these layers. Using pump wavelength of 932 nm and signal wavelength of 981 nm, a rather high net gain per unit length of 817 dB/cm is determined despite sub-optimal thermal and inversion conditions. The gain value demonstrated represents only a fraction of the theoretical gain of the material. Efficient heat dissipation from the sample is expected to improve the gain substantially.
