Neodymium-doped vanadate has become the crystal of choice for many diode-pumped lasers, primarily because the absorption cross section for 808-nm laser-diode pumping to the neodymium upper laser level is nearly eight times greater than the corresponding value in Nd:YAG. Other advantages of vanadates over the traditional YAG host are their higher emission cross sections, which result in lower CW laser thresholds, and their broader pump-light absorption bandwidths. A drawback of yttrium vanadate (Nd:YVO4
) is its low thermal conductivity, but gadolinium vanadate (Nd:GdVO4
) has thermal conductivity similar to Nd:YAG's and also has absorption and emission characteristics that are similar to Nd:YVO4
Despite these advantages, the vanadates have some shortcomings as laser hosts. Their high emission cross sections also limit their energy-storage capacity, which leads to smaller pulse energies and lower peak powers in Q-switched lasers. Moreover, the crystals are difficult to grow. Laser crystals, both YAG and vanadate, are traditionally grown by the Czochralski technique, which involves touching a seed crystal to the surface of molten material in a crucible and then slowly pulling the seed crystal upward, allowing the crystal boule to solidify beneath it. But the vanadates' high melting temperature makes this process difficult, leading to limited crucible lifetimes and inconsistent chemical composition of the crystal.
Recently, these shortcomings have been addressed by scientists in different countries, sometimes working together and sometimes independently. A collaboration between researchers in China and Germany has shown that a mixed-vanadate host (Nd:Gd0.64
) has greater energy-storage capacity and, hence, better Q-switched performance than either Nd:YVO4
. A research team in Japan has demonstrated that high-quality Nd:GdVO4
crystals can be grown by a non-Czochralski technique. And scientists in Romania and Japan have reported that ultrahigh-efficiency CW operation in Nd:GdVO4
can be obtained by pumping neodymium directly into its upper laser level.Improved Q-switching
The Chinese-German collaboration included workers from Shandong University in Jinan, China, and from the Technische Universität Berlin. The participants tested three crystal samples in a passively Q-switched resonator that was pumped by a fiber-coupled diode laser (Figure 1). Each crystal was 3.5 × 3.5 × 6 mm, with 0.5 percent atomic doping of neodymium. The first crystal was Nd:YVO4
, the second was Nd:GdVO4
and the third, the mixed crystal Nd:Gd0.64
Figure 1. The energy-storage capacities of three different laser crystals were evaluated in a passively Q-switched resonator.
They first operated each crystal in the resonator without the Q-switch, obtaining similar CW operation from all three, and thereby establishing that any difference in Q-switched performance was not caused by differences in the optical quality of the crystals. Then they added a Cr4+
:YAG saturable absorber Q-switch and compared the Q-switched performance of the three laser crystals.
The pulse-repetition frequency of a passively Q-switched laser increases with pump power because the population inversion more quickly builds to the point where spontaneous emission and amplified spontaneous emission saturate the Q-switch. But in comparing the three crystals, the researchers found that, for any given pump power, the Nd:Gd0.64
had a much lower pulse-repetition frequency than the others. That is, more energy was stored in the mixed-crystal host before the spontaneous emission bleached the Q-switch.
They also observed that, for any given pump power, the duration of the Q-switched pulse was much shorter with the Nd:Gd0.64
than with either of the other hosts. Shorter Q-switched pulses are indicative of higher stored energy because the avalanchelike emptying of a population inversion by stimulated emission occurs more quickly with a larger initial population inversion.
Figure 2. The superior energy-storage capacity of the mixed-crystal host Gd0.64V0.36VO4 enabled it to produce far more pulse energy than either YVO4 or GdVO4.
Most significantly, the scientists observed considerably higher pulse energy from the Nd:Gd0.64
than from the other two crystals (Figure 2). They attribute the improved Q-switched performance of the mixed-crystal host primarily to the inhomogeneous broadening of the fluorescence line. Just as the random local electric fields in a glass laser rod impart to Nd:glass a broad, inhomogeneous line, the random placement of Y or Gd ions in the Nd:Gd0.64
crystal accounts for an inhomogeneous broadening of the line and a corresponding decrease in the emission cross section. Indeed, when the researchers measured the fluorescence linewidth of the Nd:Gd0.64
, they found it more than three times wider than that of Nd:YVO4
The Japanese team included members from Riken (The Institute of Physical and Chemical Research) in Wako, NEC Tokin Corp. in Tsukuba and Hokkaido University in Sapporo. The researchers grew homogeneous, high-quality Nd:GdVO4
crystals with the floating-zone method, which does not require a crucible. They first mixed Gd2
powders in stoichiometrically correct quantities and then calcined the mixture at 650 °C for 10 hours in air to form single-phase Nd:GdVO4
. They sealed the calcined powder inside a rubber bag and subjected it to hydrostatic pressure of 100 MPa to create a cylindrical feed rod.
In an image furnace, they melted the bottom end of the vertical feed rod with infrared radiation from a pair of 1.5-kW halogen lamps. They pressed a seed crystal against the melted material and slowly moved the molten zone upward. The molten material, constrained by surface tension, solidified as a single crystal onto the seed crystal. The diameter of the single-crystal material was approximately 10 mm, one-third that of the feed rod (Figure 3).
Figure 3. Single-phase Nd:GdVO4 (right) is transformed into single-crystal Nd:GdVO4 (left) in the "floating zone" of molten material (center).
The researchers obtained growth rates as high as 55 mm per hour -- 10 times that typically achieved with Czochralski growth -- with the floating-zone technique. They produced defect-free crystals with neodymium atomic doping as high as 15 percent.
They tested their crystals by pumping them in planoconcave resonators pumped with both Ti:sapphire and diode lasers. They used pump radiation at 808 nm, which excited the 4
level and its neighbors, and at 879 nm, which directly excited the 4
upper laser level (Figure 4). When pumping with a laser diode at 879 nm, they observed an optical-to-optical slope efficiency of 78 percent and overall optical-to-optical efficiency in excess of 70 percent.
Figure 4. Neodymium functions as a normal four-level laser when pumped at 808 nm to one or more of several energy levels around 4F5/2. But its quantum efficiency is boosted if it is pumped with 879-nm light directly to the 4F3/2 upper laser level.
Highly efficient CW lasing
These results are nearly as good as recently reported "highly efficient" results with Czochralski-grown vanadate, and would seem to validate the floating-zone technique as a viable alternative to the problematic growth of vanadate crystals in crucibles by the Czochralski technique.
Researchers from the Institute for Atomic Physics in Bucharest, Romania, collaborated with others at the Institute for Molecular Science in Okazaki, Japan, to achieve what they believe are the highest slope efficiencies yet observed in Nd:GdVO4
. Using Ti:sapphire pumping at 879 nm, they excited Nd:GdVO4
directly into the 4
upper laser level, and they observed an optical-to-optical slope efficiency of 79 percent and overall optical-to-optical efficiency of 78 percent.
They estimate that pumping directly into the upper laser level reduces the fraction of absorbed power converted into waste heat by approximately 30 percent. The maximum possible value of slope efficiency, limited by the quantum efficiency of the 4
sequence (Figure 4), is approximately 83 percent, and the scientists calculate that the difference between this and their observed value can be accounted for by residual intracavity optical losses within the laser resonator. They obtained less optimal results when they substituted an 879-nm diode laser for the Ti:sapphire pump laser because the spatial overlap between the pump volume and the laser-mode volume was less satisfactory.