Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
D Laser Spectroscopy of Conclensed Matter INTRODUCTION The advent of lasers has greatly facilitated research in materials science. That a laser beam has the characteristics of high intensity, strong directionality, and extreme monochromaticity, and can appear in ultrashort pulse form, makes it a unique tool for materials studies. In the past two decades, a large number of laser techniques have been invented for the investigation of matter in all its phases. They have opened the possibility of research in many hitherto unexplored areas of materials science. Generally speaking, lasers can be applied to two types of problem: (1) to probe a material and (2) to modify or process a material. The past decade has witnessed great advances along both lines. For example, laser light scattering has become a conventional technique for studying excitations in condensed matter; nonlinear optical spectroscopy allows the study of forbidden transitions in a medium and the study of homogeneous broadening of spectral lines as narrow as PI kHz; transient optical spectroscopy can probe dynamic properties of a medium on a time scale as short as a subpicosecond (i.e., <10-'' second); optical mixing is useful for monitoring and studying molecular adsorbates on surfaces; and laser heating is promising as a new method for annealing crystalline films or for growing various types of amor 258
APPENDIX D 259 phous, crystalline, and alloy layers. As laser techniques become increasingly more sophisticated, we anticipate in the coming years an exciting period for laser-related materials research, from both the practical and scientific points of view. ACCOMPLISHMENTS OF THE PAST DECADE Spectroscopic study is essential for the microscopic characterization of a medium. Laser spectroscopy has brought new life to optical spectroscopy of condensed matter. In this section we survey the accomplishments of the past decade. Nonlinear Optical Spectroscopy This type of spectroscopy flourishes only because tunable lasers have become easily available. It comes in many different forms, depending on the nonlinear optical process involved. A few of them are considered here. TWO-PHOTON SPECTROSCOPY Two-photon spectroscopy is commonly used to study transitions between states of the same parity. In research on excitor polaritons in a semiconductor, for example, since two-photon spectroscopy is unaffected by the reststrahlen band it can yield detailed information about the damping and dispersion curve of the excitor polaritons. The technique has also been applied to the study of excitonic molecules in solids, which is a subject of immense theoretical interest. More recently, two-photon spectroscopy has been used to measure the (din (off transitions of rare-earth ions in solids not observable in one-photon absorption owing to the presence of the (din ~ In- transitions. This is most interesting because through such an investi- gation one can expect a much better understanding of rare-earth ions and their interaction with the lattice in a solid. HOLE BURNING IN INHOMOGENEOUSLY BROADENED SPECTRA A laser beam can be intense enough to saturate a transition in a medium. If the laser linewidth is much narrower than the inhomogene- ous broadening of the line then, with the laser frequency fixed, only a small group of ions or molecules can be resonantly excited by the laser. By saturating the transition of this group of ions or molecules, a hole is
260 APPENDIX D created in the inhomogeneously broadened spectrum. The hole width is generally limited by homogeneous broadening, assuming that the laser linewidth is negligible. This hole-burning effect makes high-resolution spectroscopy of ions or molecules in condensed matter a reality. Both absorption or fluorescence can be employed to probe the holes. Lines as narrow as a few megahertz have been observed. This saturation- spectroscopy technique can be used to study hyperfine and superhy- perfine interactions of rare-earth ions in solids and fine structure of organic molecules in solid matrices. At sufficiently low temperatures, the laser-induced holes can last almost indefinitely. As many as ~105 holes can be burned on an inhomogeneously broadened line 10 cm- wide. They can therefore be used for making optical memory devices with high densities of data bits. Such a possibility is currently being pursued vigorously by several industrial laboratories. OTHER NONLINEAR SPECTROSCOPY TECHNIQUES Four-wave mixing and coherent Raman spectroscopy allow us to study excitations in both the ultraviolet and the infrared ranges. Because their sensitivity is high, they can be employed to detect impurities in condensed matter. Their spectral resolution is limited only by the laser linewidth. Because the output is coherent and directional, spatial filtering can be adopted to suppress the lumines- cence background. Thus, these techniques can be used to study excitations that are normally masked by luminescence. Applications of these techniques to condensed-matter physics have attracted only limited attention in the past; but more recently, with the advance of laser technology, they have begun to receive increasing recognition in the community. Transient Optical Spectroscopy Transient coherent phenomena arising from the resonant interaction of radiation with matter are among the most fascinating topics in condensed-matter physics. They were studied extensively in magnetic resonance before the laser era. With lasers, extension of these studies to optical transitions becomes possible. The past decade has seen increasing activity in this area of research. With the use of photon echo and optical notation techniques, for example, homogeneous linewidths of optical transitions of rare-earth ions in crystals as narrow as ~ 1 kHz have been measured. Thus, detailed information about spin-spin inter- actions between the ions and surrounding atoms can be obtained.
APPEND/X D 261 Ultrafast Laser Spectroscopy PICOSECOND LASER SPECTROSCOPY In 1966, laser scientists discovered how to use the technique of mode locking to produce optical pulses as short as ~5 ps in duration. Stimulated by progress in laser technology, the past decade has seen dramatic improvements in picosecond instrumentation and techniques, with the result that picosecond laser spectroscopy is an area experi- encing major advances at present. Several laser manufacturers now sell reliable cw mode-locked dye lasers delivering ~3-picosecond-long pulses at ~100-MHz repetition rates. These pulses may be selectively amplified to energies in the millijoule regime (i.e., having peak powers of ~108 watts), and pulse length measurements, including delay times, are routinely carried out by the technique of autocorrelation by . . . non Inear mixing. In general, by exciting a medium with a picosecond pulse, followed by probing with another picosecond pulse, the dynamic properties of the medium can be studied on a picosecond time scale. This is exciting since it offers the opportunity to measure directly, for the first time, the carrier relaxation time, the excitation lifetime, the phonon relaxation, and other properties in a condensed medium. It opens a new, important area of research that seldom has been explored in the past. As an example, time-resolved photoemission spectroscopy (PES) of semiconductor, metal, and insulator surfaces yields important data on the transient behavior of selectively excited carriers. Using a picosec- ond laser pulse to excite a narrow energy distribution of electrons, a delayed picosecond pulse has been used to probe the relaxed distribu- tion by PES. By energy analyzing the photoemitted electrons as a function of time delay, fundamental information about the energy relaxation processes affecting the electron distribution was obtained. Furthermore, by using circularly polarized laser light and studying the spin polarization of the photoemitted electrons, phase-destroying pro- cesses may be studied. Thus momentum relaxation processes may be distinguished from energy relaxation processes. In a new approach to high-speed electronic instrumentation, known as picosecond optical electronics, a short laser pulse illuminates a high-speed photoconductor, thereby producing a fast switch for an electrical signal. By combining two of these switches with a control- lable time delay, a sampling system capable of time resolution of better than 2 picoseconds has been demonstrated. Studies of propagation delays between the drain and gate signals of GaAs ferroelectric
262 APPENDIX D transistors have been carried out. The response time of a high-speed silicon photodiode has been measured. Electron transport in materials with particular types of electronic defects has been characterized with picosecond time-resolved photocurrent spectroscopy. The extension of optical electronics to the subpicosecond regime calls for ways of overcoming dispersive and capacitive effects in electronics compo- nents. Efforts are being devoted to finding better ways of rectifying the light pulses to produce short dc pulses (such pulses really look like microwave radiation since they contain frequencies in the terahertz range). FEMTOSECOND LASER SPECTROSCOPY In several research laboratories techniques were developed recently to produce laser pulses as short as 70 femtoseconds (1 femtosecond = 10-'5 second). A method of pulse compression has recently been developed that has resulted in the realization of pulses only 16 femtoseconds long. One of the results that has been achieved by the use of such ultrashort laser pulses is described in what follows. An area of great interest over the past several years has been laser processing and laser annealing of materials. While many different facets of this field have been explored by studying the results of laser irradiation of materials, little has been done to time-resolve the actual annealing process. Recently, such studies have been carried out with femtosecond pulses, and the results support a model where the light is absorbed and first creates electron-hole pairs, after which the irradi- ated surface is converted from a crystalline structure within an electron-hole plasma to a molten state. By such studies, one can determine the dynamics of the energy absorption process responsible for laser processing. This is an exciting new area of materials science. SOME DIRECTIONS FOR FUTURE RESEARCH Now that laser technology has become more mature, one can anticipate rapid growth in several areas of laser applications to condensed-matter physics in the coming years. The increased use of two-photon spectroscopy for the study of forbidden transitions of rare-earth ions in solids is anticipated. In studies of materials properties, high-resolution laser spectroscopy and nonlinear optical spectroscopy are expected to become common laboratory techniques. The future of transient optical spectroscopy for the study of optical
APPENDIX D 263 transitions of rare-earth ions in crystals is particularly bright because these techniques can be extended to the study of transitions between excited states. Many of the sophisticated techniques developed earlier in magnetic resonance are yet to be transplanted to this field. Another area of rapid growth in the laser spectroscopy of condensed matter is laser studies of surfaces. Both laser perturbation and laser probing of surfaces are exciting new areas of research that have hardly been explored. It is clear that we have entered a new era in making measurements on subpicosecond time scales. Laser-based techniques will allow such measurements to be made routinely in physics, chemistry, biology, materials science, and device studies. This area of research is now known as femtosecond science. Some of the interesting subjects for study in this area are listed below. By using a femtosecond laser pulse to create photoemitted electrons with a small energy spread and by accelerating these electrons across a large potential, a pulse of electrons is made available for time- resolved electron-diffraction studies. For the first time, such structural changes as melting and structural phase transitions may be studied on the femtosecond scale. The response of crystalline solids to a short electrical pulse, i.e., the electric field of the laser pulse, may be studied by femtosecond spectroscopy. For example, the Franz-Keldysh effect (the change of the band gap of a semiconductor in an electric field) can be studied in a time-resolved fashion. When a femtosecond light pulse causes photoemission, the electron pulse and the light pulse are synchronized. This makes it possible to carry out studies where one pulse creates an excitation and the other pulse probes this excitation. Thus the electron pulse may create a change in matter than can be monitored optically by the light pulse. It might be possible to create or disturb a solid-state plasma using the short electron pulse and then probe it by time-resolved reflectivity measurements. In the area of femtosecond-pulse technology, research is aimed not only at producing still shorter pulses but also at developing techniques to extend the wavelength region of coherent-light generation toward the ultraviolet and the infrared regions. On the side of laser technology we can also expect to see advances in several areas. A severe limitation in the progress of high-power laser technology comes from optical breakdown. This is a subject of great practical importance and has been pursued vigorously in the past two decades. From the scientific point of view, this is also a subject of great
264 APPENDIX D interest. Optical breakdown in solids is a highly nonlinear process. How a laser beam excites the carriers, generates the plasmas, and eventually shatters or melts a solid is not a trivial matter to understand. Breakdown mechanisms for excitations with different laser frequencies and pulse widths could be different. Solution of the problem requires the joint efforts of theorists and experimentalists. Progress in this area is being made step by step. It is hoped that in the next decade, optical materials of much improved quality will be produced as a result of the continuing research effort. Another important area of materials research related to laser tech- nology is the search for better nonlinear optical crystals. Such crystals are essential in the extension of coherent light sources to the ultraviolet and infrared regions. Recent theoretical calculations are fairly success- ful in predicting nonlinearities of certain crystals. Organic and inor- ganic crystals of large nonlinearities have actually been grown follow- ing the theoretical hints. This is encouraging. It is likely that new and better nonlinear optical crystals suitable for efficient frequency con- version over a broad range and for use in nonlinear optical devices for data processing will emerge in the near future. Optical fibers have grown, in the past decade, into an important branch of the optical industry. They are the key element for future optical communications and data processing. An exciting development in the research into fiber materials has taken place within the past year. It is found that crystals can also be pulled into thin optical fiber form. Thus, fiber lasers and fibers for optical frequency conversion may soon appear in laboratories. Their possible applications are numerous, limited only by imagination. They are likely to revolutionize science and technology in many disciplines. The prospect of this field is truly great. Its progress in the next 10 years will be interesting to follow.