Solid-state materials tune across the spectrum
Paul M. W. French
Traditionally dye lasers have been employed where tunability was required. Due to such drawbacks as the often user-hostile nature of the technology, the associated safety issues, and the relatively high level of expertise required to run and maintain high-performance dye-laser systems, they are increasingly being replaced by solid-state lasers. While the trend toward tunable solid-state sources was hastened by the development of tunable flashlamp-pumped alexandrite systems, it was the demonstration of the Ti:sapphire laser by Peter Moulton that heralded the revolution in solid-state laser technology and application. The medium was essentially compatible with existing dye-laser cavities and pump sources, making conversion from a liquid to a solid-state system almost as easy as swapping a dye cell for a laser rod.
The success of the continuous-wave (CW) Ti:sapphire laser illustrated the potential benefits of laser pumping, rather than lamp pumping, of solid-state media. Known broadband laser media became much more attractive when pumped with commercially available, high-spectral-brightness noble-ion, neodymium-doped yttrium aluminum garnet (Nd:YAG) and semiconductor lasers rather than with flashlamps. For many types of laser crystals, it is more straightforward and efficient to directly pump the absorption bands and often possible to select a pump wavelength that avoids problems of excited-state absorption (see table on p. 94).
A second factor driving the field forward was the impressive creativity of the laser crystal growth community. Scientists in this field can now develop "designer" laser crystals for particular applications and have developed, or rediscovered, a number of vibronic media for tunable solid-state lasers. Important solid-state broadband laser media provide spectral coverage from the ultraviolet to the near-infrared (NIR) spectral regions (see Fig. 1).1-3 Second-harmonic generation is routinely achieved with up to tens of percents of efficiency, providing further coverage at wavelengths below approximately 500 nm, including the ultraviolet region. Coverage is most sparse in the visible spectral region around 600 nm where, ironically, the laser dyes are at their best. Notable omissions from the summary of tunable solid-state materials include the relatively nontunable Nd3+-, Er3+-, and Ho3+-doped laser media that operate at NIR wavelengths.
The spectral region from roughly 650 to 1100 nm is well served by Ti:sapphire, which is an almost perfect CW laser medium except that its absorption bands are not appropriate for pumping with currently available high-power diodes and are not well suited for flashlamp-pumping. The issue of pumping with more-convenient sources is addressed by the chromium-doped colquiriite crystals lithium calcium aluminum fluoride (Cr:LiCAF),4 lithium strontium aluminum fluoride (Cr:LiSAF),5 and lithium strontium gallium aluminum fluoride (Cr:LiSGAF),6 originally developed at Lawrence Livermore National Laboratory (Livermore, CA).
The chromium-doped colquiriites provide a spectral coverage similar to Ti:sapphire but have broad absorption bands suitable for pumping with flashlamps or with diode lasers operating near 670 nm. Furthermore, their longer upper-state lifetimes make them more useful as amplifier media. Another chromium compound, alexandrite (Cr3+:BeAl2O4), can also be pumped by flashlamps or diode lasers but had enjoyed limited application due to thermal problems and excited-state absorption.7 These lasers have numerous scientific applications. The coincidence of this spectral region with the absorption minimum of biological tissue has recently prompted interest from the medical diagnostic/imaging community.
Chromium-doped media also cover the NIR spectral regions, with the Cr4+ ion providing the requisite energy levels. For example, chromium-doped forsterite (Cr4+:Mg2SiO4) provides room-temperature CW tuning ranging from roughly 1.21 to 1.32 µm, with pulsed operation covering the region from 1.13 to 1.37 µm. The CW tuning range can be enhanced by cryogenic operation. Room-temperature CW output tunable from 1.37 to 1.58 µm is available from Cr4+:YAG. Developed in Japan/USA and Russia, respectively,8,9 these media have laser parameters comparable to those of Ti:sapphire. They should provide comparable performance in their respective spectral regions, although there are significant problems to be addressed concerning thermal lensing.
Tunability around 1.3 µm and 1.55 µm is important for telecommunications applications, and 1.54 µm is an eye-safe wavelength of major importance for optical ranging and other applications. The second harmonic of Cr:forsterite covers the red spectral region and may prove important for applications such as photodynamic therapy and two-photon microscopy. Both of these chromium-doped media could provide attractive pump lasers for optical parametric generation in the mid-infrared region.
At shorter wavelengths, ytterbium-doped lasers such as Yb:YAG, Yb:Ca5(PO4)3F, and Yb:BaCaBO3F provide limited CW tunability around 1.04 µm and appear to be promising candidates for high-power lasers and pump sources for other solid-state laser media.10,11
Yet further into the infrared, tunable gain is provided by Cr2+ ions doped in II-VI compounds.12 Pulsed room-temperature laser action has been demonstrated from 2.3 to 2.5 µm, and the prospects for CW laser action in this spectral region look promising. At present the principal laser medium for this spectral region is cobalt-doped magnesium fluoride (Co2+:MgF2), which provides pulsed room-temperature operation and CW tuning from 1.62 to 2.1 µm at 77 K.13 Crystals doped with thulium (Tm3+), such as Tm3+:YAG, Tm3+:YLF, and Tm3+:yttrium scandium gallium garnet (Tm3+:YSGG), also provide tunable gain in the near-infrared and CW operation at room temperature.14,15 This spectral region could be important for molecular spectroscopy with particular practical relevance to remote sensing and pollution monitoring.
There are numerous scientific applications of visible radiation, particularly in chemistry, biology, and medicine, but this spectral region is currently not well served by solid-state lasers. There is no broadband vibronic visible laser, and even the second harmonics of NIR lasers provide only limited coverage. Thus the dye laser survives as the most viable tunable visible laser with competition only from complex and expensive ultraviolet-pumped optical parametric generation systems. Polymer-based solid-state dyes have been developed for users who cannot tolerate the inconvenience of the liquid state. These media are suitable for pulsed but not CW operation.
Another alternative is the praseodymium-doped yttrium lithium fluoride (Pr:YLF) laser, which provides CW room-temperature operation at a number of discrete wavelengths in the visible and NIR spectral regions from 522 to 904 nm.16 This laser can provide red, yellow, and green output and, together with a blue pump source, yields an RGB laser system that may find application in printing and display technology (see photo on p. 93).
The ultraviolet spectral region is well covered by harmonic generation of NIR solid-state lasers. In addition, researchers can consider a growing number of cerium-doped lasers that tune around 300 nm. Cerium-doped colquiriites such as Ce3+:LiCAF17 and Ce3+:LiSAF18 can be pumped at ~266 nm, for example, by a frequency-quadrupled Nd:YAG laser and potentially provide gain from ~280 to 320 nm. Ce3+:YLF and cerium-doped lutetium lithium fluoride (Ce3+:LuLiF6) can be pumped at 248 nm by a krypton fluoride excimer laser to cover the region from about 305 to 335 nm.19,20 Continuous-wave laser operation has not yet been demonstrated but appears feasible.
Tunable/ultrafast solid-state laser cavities
Solid-state lasers are typically pumped transversely by flash lamps or arc lamps or longitudinally by another laser. The initial commercial Ti:sapphire laser cavities looked remarkably similar to dye laser cavities; the ease of substituting the solid-state laser crystal for an existing dye cell in the output of an argon-ion-laser pump source is one reason this field has progressed so rapidly.
In a typical cavity, curved folding mirrors focus the intracavity radiation into the gain medium to achieve the high power density necessary for CW pumping (see Fig. 2). For the optimum folding angle, the astigmatism introduced by the off-axis incidence at the curved mirrors is balanced by the astigmatism introduced by the Brewster-angled laser rod located at the cavity beam waist. Typical beam-waist radii are 20 to 50 µm, adjusted such that the confocal parameter is matched to the length of the rod. This dimension is determined by the doping level achievable and the corresponding length required to absorb most of the pump power. Active-ion doping levels generally range from about 0.2% to 20%, and rod lengths range from 1 to 20 mm. With TEM00 diffraction-limited pump beams, available from noble ion or Nd:YAG lasers, mode-matching is straightforward. Component X might be a frequency-selective element for tunable operation, a Pockels cell for Q-switched operation, a modulator for picosecond modelocked operation or a prism pair and spatial filter for femtosecond operation.
This cavity geometry had a largely unexpected consequence. Due to the relatively large nonlinear interaction length caused by propagation of the focused cavity radiation through about 1 cm of gain medium, the nonlinear phase change due to the intensity-dependent refractive index was much larger than had been observed with other gain media such as CW dye lasers. This first resulted in the generation of anomalously short pulses from conventionally modelocked solid-state lasers and later resulted in the experimental discovery of what is now known as Kerr lens modelocking (KLM).21 This modelocking technique has been the subject of intensive investigation and is now a reasonably well-understood method of generating femtosecond pulses significantly shorter than 20 fs.22,23
Self-starting Kerr lens modelocking
Essentially picosecond and femtosecond pulses can be generated by any solid-state laser with sufficient gain linewidth to support the desired pulse durations. Recently Cerullo and others described how to achieve self-starting KLM in a Ti:sapphire laser.24 Thus femtosecond pulses are available from a solid-state laser cavity containing only a gain medium, a device for controlling intracavity group velocity dispersion (GVD) such as a prism pair, and a spatial filter, typically a slit.
At Imperial College, we have demonstrated self-starting KLM in Ti:sapphire, Cr:YAG, and Pr:YLF lasers and are confident that, given sufficient mode-matched pump power, usually less than 4 W, any solid-state laser can exhibit self-starting KLM operation. This observation suggests that reliable and user-friendly ultrafast solid-state lasers could be also be low-cost--thereby making them attractive for a wide range of applications--but for the requirement of the pump laser. The ultrafast community has addressed this last issue with the demonstration of pumping with diode-laser sources.
One consequence of the increased convenience of solid-state laser technology is that user-friendly commercial ultrafast laser systems now routinely outperform "home-made" research lasers. This is driving forward developments in user communities such as chemistry, biology, and optical signal processing. The availability of compact, reliable, all-solid-state diode-pumped lasers tuning throughout the visible and infrared spectral regions will be important for metrology, remote sensing, environmental monitoring, medical diagnostics, communications, data storage, display technology, and many other applications. In addition to shrinking system sizes from several meters down to tens of centimeters, these improved sources should offer sufficient purchase and maintenance cost savings to create significant new markets.
Progress in all-solid-state laser technology continues to be extremely rapid. Compact, diode-pumped, all-solid-state lasers delivering tunable narrow linewidth operation, both CW and Q-switched, as well as tunable picosecond and femtosecond operation, have been demonstrated. The biggest challenge in designing diode-pumped solid-state lasers is accommodating the much lower brightness of the diode pump lasers. Although high-power, diffraction-limited beams are now available from relatively expensive tapered amplifier structures, the spatial beam profiles of most diode lasers and high-power arrays are usually multimode and far from diffraction limited.
Current techniques include end-pumping, which entails shaping the diode-pump beam to mode-match a cavity beam waist as closely as possible, analogous to traditional laser-pumping. For higher-power lasers, side-pumping with diode arrays in a manner similar to flashlamp-pumping has produced up to kilowatts of output power. Both approaches can yield high-quality output beams from the all-solid-state laser that can match those from their large-frame ancestors. It is not unreasonable to expect that within a few years diode-pumped system performance will nearly equal that of large-frame systems. If this prospect is not realised in specific cases, it will probably not be because it is not technologically possible but because there is insufficient reason to do it.
Future developments will be driven by the availability of diode pump lasers and suitable laser media. Most of the laser media represented in Fig. 1 can be diode-pumped directly or can be pumped by another diode-pumped all-solid-state laser. The cost and availability of appropriate pump diodes will mainly be determined by demand; mass-market applications are probably necessary to achieve the cost reductions necessary for widespread distribution of all-solid-state lasers. The situation will be helped by the development of laser media designed to better exploit existing high-power/low-cost diode technology. For the scientific community, the trend will continue toward solid-state laser systems and therefore toward more user-friendly, compact, and probably less expensive sources. n
REFERENCES
1. W. Koechner, Solid-state Laser Engineering, 2nd ed., Vol. 1 of Springer series in optical sciences, Springer-Verlag, Berlin, Germany (1977).
2. S. A. Payne and G. F. Albrecht, "Solid-state Lasers," in Encyclopedia of Lasers and Optical Technology, 603 (1987).
3. P. F. Moulton, Proc. IEEE 80, 348, (1992).
4. S. A. Payne et al., IEEE J. Quant. Electron. QE-24, 2243 (1988).
5. S. A. Payne et al., J. Appl. Phys. 66, 1051 (1989).
6. L. K. Smith et al., IEEE J. Quant. Electron. QE-28, 2612 (1992).
7. J. C. Walling et al., IEEE J. Quant. Electron. QE-21, 1568 (1985).
8. V. Petricevic, S. K. Gayen and R. R. Alfano, Appl. Phys. Lett. 52, 1040 (1988).
9. N. B. Angeert et al., Sov. J. Quantum Electron. 18, 73 (1988).
10. L. D. DeLoach et al., IEEE J. Quant. Electron. QE-29, 1179 (1993).
11. U. Brauch et al., Opt. Lett. 20, 713 (1995).
12. R. H. Page et al., CLEO `95, Baltimore, MD, paper WH5 (1995).
13. D. Welford and P. F. Moulton, Opt. Lett. 13, 975 (1988).
14. J. R. C. Stoneman and L. Esterowitz, Opt. Lett. 15, 486 (1990).
15. J. F. Pinto, L. Esterowitz and G. H. Rosenblatt, Opt. Lett. 19, 883 (1994).
16. T. Sandrok et al., Appl. Phys. B 58, 149 (1994).
17. M. A. Dubinskii et al., J. Mod. Opt. 40, 1 (1993).
18. J. F. Pinto et al., Electron. Lett. 30, 240 (1994).
19. D. J. Ehrlich, P. F. Moulton, and R. M. Osgood Jr., Opt. Lett. 4, 184 (1979).
20. N. Sarukura et al., Opt. Lett. 20, 599 (1995).
21. D. E. Spence, P. N. Kean, and W. Sibbett, CLEO `90, Anaheim, CA, paper #CPDP10-1 (1990).
22. F. Krausz et al., IEEE J. Quant Electron. QE-28, 2097 (1992).
23. P. M. W. French, Rep. Prog. Phys. 58, 169 (1995).
24. G. Cerullo, S. De Silvestri, and V. Magni, Opt. Lett. 19, 1040 (1994).
Jason Sutherland of Imperial College demonstrates simultaneous lasing of Pr:YLF at 522 and 604 nm. When 476-nm output from the argon-ion-laser pump beam is included, the resultant RGB system is suitable for color projection.
FIGURE 1. Tunable solid-state lasers provide coverage across much of the near-infrared spectral region and well into the ultraviolet. Second-harmonic generation from infrared sources (not shown) increases available output at higher frequencies.
FIGURE 2. Typical solid-state laser cavity includes collimated region for inclusion of intracavity element (Component X) such as a Pockels cell, a modulator, or a prism pair.