The elements of the lanthanide series, from Ce (atomic number 58) to Yb (atomic number 70), form a group of chemically similar elements that have in common a partially filled 4f shell. These so-called “rare earth” (RE) elements usually take on a 3+ ionic state (RE3+). Because the 4f electronic-energy levels of each lanthanide ion are shielded from external fields by 5s2 and 5p6 outer-shell electrons, RE3+ energy levels are predominantly independent of their surroundings.

The characteristic energy levels of 4f electrons of the trivalent RE elements have been investigated in detail by Gerhard Heinrich Dieke and co-workers and were reported approximately 30 years ago. The Dieke diagram showing RE3+ energy levels is a familiar tool of scientists and engineers working with RE elements. However, the history of RE elements goes back to the year 1787 in the small Swedish town of Ytterby near Stockholm and to the gifted amateur mineralogist and military man Lt. Carl Axel Arrhenius. Arrhenius discovered an unusual black mineral in Ytterby (perceived initially as much rarer in occurrence and in concentration than the common ores or earths of aluminum, calcium, etc.). Many new elements were discovered by various chemists upon analysis of this black stone and others like it. The names given to these elements are variations of the location where the first discovery was made: yttrium, ytterbium, terbium, and erbium. The history of RE elements is fascinating and involves many other famous names in science: Berzelius, Gadolin, Bunsen.

The properties of these elements and their multifaceted applications to science and industry are equally fascinating and have remained important to this day. Commercial applications of RE elements began after World War II, when their available quantity and purity were greatly enhanced by improved separation techniques developed as a part of the Manhattan Project. Until fairly recently, the main industrial application of RE elements has been in permanent magnets. The unpaired 4f electrons result in some RE elements having the highest magnetic moments of any element. The development and applications of RE magnets are reviewed in a very interesting article by Livingston3 in a previous MRS Bulletin issue. In this issue of MRS Bulletin, we have taken as our aim to review some of the properties and applications of RE elements relevant to photonics.

An issue on novel applications of materials doped with rare-earth (RE) ions can scarcely fail to address lasers, but it need not address all RE-based lasers. Some Nd3+ -doped lasers, particularly Nd:YAG (Y3Al5O12, yttrium aluminum garnet), emitting light with a wavelength of 1064 nm, are very well-established commercial products—by no means novelties.1 Some other near-infrared (NIR) lasers, based on Er3+ or Tm3+, are also available commercially. That wavelength region is relatively easy for RE laser ions, involving energy spacings between initial and final energy levels small enough to give large stimulated emission cross sections for useful, long upper-state life-times, yet large enough to minimize thermal deexcitation mechanisms. On the other hand, RE-doped lasers for ultraviolet (UV) and visible wavelengths are quite novel, since efficient laser operation is more difficult to achieve in these spectral ranges. Intriguing progress on such devices has been made in recent years, driven by several important applications.

In this article, we begin by noting some of the alternative ways to obtain laser light at these wavelengths, including their advantages and drawbacks. We then discuss basic properties of RE-doped laser materials and how these can be advantageous. We then review a few of the most important and recent RE-doped laser materials and techniques for obtaining UV and visible output.

Semiconductors doped with rare-earth (RE) elements have attracted a lot of attention as alternative materials for producing electrically pumpe d semiconductor lasers whose emission wavelength is very weakly dependent on temperature. This prospect is especially attractive in the case of indirect-gap Silicon, whose photonic applications as the material for light emitters still remain more of a hope than a reality. In view of a desirable emission wavelength at 1.5 μm, a lot of research has concentrated on Si:Er (see Coffa et al. for a recent review). It is generally recognized that doping with Er ions presents one of the most promising approaches to Silicon photonics. However, despiteintensive investigations, stimulated emission has not been conclusively demonstrated for Si.Er or for any other RE-doped semiconductor. This is in striking contrast to optical amplifiers and lasers based on various erbium-doped glasses. In this article, which builds on recent articles in MRS Bulletin on Silicon photonics, we will address the issues relevant to efficient light generation by semiconductors doped with RE elements in general, and specifically by Si:Er-based structures.

The intraimpurity electronic structure of RE ions is dominate d by electron-electron and spin-orbit interactions within the 4f shell. In the case of Er3+, they produce separated J-multiplets with 4I15/2 and 4I13/2 as the ground and the lowest-lying excited states, respectively. Due to the effective Screening of 4f electrons by the outer electron Shells, the host has a very limited influence and changes only slightly the relative positions of the levels. Depending on a particular site symmetry, the even terms of the crystal field split the free-ion J-multiplets into the Stark components typically by several meV for the ground State. The energy-level diagram of an Er3+ ion in a cubic crystal field is shown in Figure 1, where the energy transfer paths relevant for Si:Er are also schematically indicated. The odd terms of the crystal field potential admix the states of opposite parity to the 4f11 configuration of the Er3+ ion, thereby introducing a certain degree of electric-dipole strength into the otherwise forbidden intra-4f-shell transitions. This effect enhance s slightly the magnetic-dipole strength of the 4I15/2 ↔ 4I13/2 transition and is host- and site-dependent. There-fore, Er-related center s of different microstructure can be fairly easily identified.

As discussed in the accompanying articles in this issue of MRS Bulletin, the optical properties of rare-earth (RE) elements have led to many important photonic applications, including solid-state lasers, components for telecommunications (optical-fiber amplifiers, fiber lasers), optical storage devices, and displays. In most of these applications, the host materials for the RE elements are various forms of oxide and nonoxide glasses. The emission can occur at visible or infrared (IR) wavelengths, depending on the electronic transitions of the selected RE element and the excitation mechanism. Until recently, the study of semiconductors doped with RE elements such as Pr and Er has concentrated primarily on the lowest excited state as an optically active transition. The presence of transitions at IR wavelengths (1.3 and 1.54 μm) that are coincident with minima in the optical dispersion and the loss of silica-based glass fibers utilized in telecommunications, combined with the prospect of integration with semiconductor device technology, has sparked considerable interest.

The status and prospects of obtaining stimulated emission in Si:Er are reviewed by Gregorkiewicz and Langer in this issue and by Coffa et al. in a previous MRS Bulletin issue. While great progress is being made in enhancing the emission intensity of Er-doped Si, it still experiences significant loss in luminescence efficiency at room temperature, as compared with low temperatures. This thermal quenching was shown by Favennec et al. to de crease with the bandgap energy of the semiconductor. Hence wide-bandgap semiconductors (WBGSs) are attractive candidates for investigation as hosts for RE doping.

Primarily to meet the dramatic increase in Internet traffic, a substantial expansion in new fiber-optic networks has been seen in the last few years, increasing the total amount of transmission fiber deployed in the field. However, increased capacity has also been achieved by utilizing more of the available bandwidth present in the currently installed fiber. A key component in facilitating this increase in bandwidth is the erbium-doped fiber amplifier (EDFA), which provides efficient broad-band gain in the 1530–1560-nm telecommunications window. Erbium-doped glasses can be drawn into low-loss fiber, and the width of the gain band can be controlled with glass composition. With appropriate composition and design, EDFAs can simultaneously amplify 32 or more wavelengths, providing a 32-fold increase in data capacity over single-channel systems. These devices can boost signal strength by a factor of 1000, with high reliability and low noise at data rates exceeding 1 Tbit/s. In this article, we review some of the properties that are key to the success of EDFAs and discuss the potential for other rare-earth-doped glass-fiber combinations that may find possible applications in future telecommunications networks.

Fiber optics have revolutionized the telecommunications industry, providing more information capacity and greater distances between signal boosters than copper wire and coaxial cable. The attenuation in coaxial systems increases exponentially with signal frequency, making high-speed transmission over long distances impractical. The best copper systems have a bandwidth of about 10 Mbit/s and are limited to lengths of less than 200 m at high data rates. In contrast, the attenuation of SiO2 optical fibers is low and independent of signal frequency, thus optical fiber can easily support 100 Gbit/s (10,000 times the capacity of copper) over 80 km and is currently only limited by the speed of the transmission and receiving electronics, with capacities in excess of 50 Tbit/s theoretically possible.1 For links in excess of 80 km, signal amplification is necessary to prevent total loss of the signal. In the 1980s, amplification was done with electronic devices called repeaters that detected the light, converted it to an electronic signal, amplified, retimed, and then retransmitted it as an optical pulse.

The field of optical telecommunications has itself undergone a revolution. In the late 1980s, the invention of the all-optical amplifier allowed for simultaneous amplification of multiple channels in a single optical fiber each at a different wavelength or color of light. SiO2 fibers have a minimum in attenuation in the infrared (IR) portion of the optical spectrum near 1550 nm, as shown in Figure 1. The EDFA fortuitously provides high gain and low noise in the 1530-1560-nm spectral window. This technology now enables simultaneous amplification of 32 channels in a single fiber without the need for optical-to-electronic conversion. Thus single-fiber capacities of 320 Gbit/s are currently being deployed today. To perform this electronically, each channel would have to be separated (demultiplexed), amplified by its own costly repeater, and then recombined (multiplexed) in the fiber. Researchers are now perfecting 100-channel EDFAs in the lab.

Conventional optical data-storage techniques, such as magneto-optic disks and CD-ROMs, record a single bit of information at each particular substrate location. In order to produce the gigabyte-class storage substrates demanded by today's computers using such conventional technologies, access to tens of billions of individual material locations is required. This brute-force approach to optical data storage has produced impressive results. However, there is increasing interest in methods for more efficiently accessing storage materials. One approach is to record multiple bits at a single storage-material location. This can be accomplished by multiplexing the bits spectrally, using differing optical frequencies to record data bits. It has been realized for over 20 years that when certain materials are cooled to appropriate temperatures, typically below 20 K, the possibility of spectrally multiplexing large numbers of bits in a single material location arises. Although this approach, known as spectral hole-burning, has been proposed as a data-storage mechanism, to date it has primarily been used as a tool to study material properties. Rare-earth-doped crystals have been demonstrated to have properties that lend themselves to a variety of different spectral hole-burning-based data-storage applications. In this article, we will review the principles of spectral hole-burning, discuss some specific material systems in which spectral hole-burning is of particular interest, and describe methods for producing high-capacity, high-data-rate spectral memories.

Spectral hole-burning, and spectral memories based on spectral hole-burning, depend on a material property referred to as inhomogeneous absorption line broadening. Materials exhibiting this property contain active atoms or molecules that individually respond to (absorb) very specific frequencies of light, but the collective response of all of the material's active atoms or molecules covers a spectral region that is broad compared with the response of a particular active atom or molecule. Inhomogeneous absorption line broadening is caused by local variations in the structure of the host, which in turn lead to variations in the electronic levels of the active atoms or molecules. The absorption linewidth of an individual absorber is referred to as the homogeneous linewidth Γh, and the absorption width of a collection of inhomogeneously broadened absorption centers is referred to as the inhomogeneous linewidth Γi. Application of monochromatic light to such a material has the effect of exciting only a very small subset of active absorbing atoms—those residing in the illuminated spatial volume within a homogeneous width of the exciting light's specific frequency. If the frequency of the imposed light is shifted, a different subset of active absorbing atoms in the illuminated volume responds.

The human eye places remarkably stringent requirements on the devices we use to illuminate objects or generate images. Exceedingly small deviations in color or contrast from what we consider natural are easily judged by the brain to be fake. Such cognition drives consumer practice, so great efforts have been made for over a century to synthesize emissive materials that match the response functions associated with the human perception of color. This is an extremely difficult task, given the diverse range of considerations, some of which include whether (1) the display is viewed under artificial light or natural sunlight, (2) the images are stationary or moving, and (3) the rendering of depth in a two-dimensional image is believable.

Established technologies including cathode-ray tubes (CRTs), vacuum fluorescent displays (VFDs), lamps, and x-ray phosphors have made possible a wide variety of display and imaging devices. However, continued advances are required to increase brightness, contrast, color purity, resolution, lifetime, and viewing angle while still lessening the cost, weight, volume, and power consumption. Mature or emerging technologies that address these issues include thin-film electroluminescent (TFEL) displays, liquid-crystal displays (LCDs),8 field-emission displays (FEDs),9 and plasma displays (PDs).10-12 Each of these technologies uses luminescent materials consisting typically of an activator from which light is emitted and a host for low concentrations of the activator (typically >1% activator). The requirements of the host and activator are discussed in a later section. The luminescent material can exhibit either a narrow emission spectrum, useful for color displays, or a broadband emission, which can extend into multiple colors. In addition, with multiple activator/host combinations, a luminescent material can emit several colors and even white light. While LCDs are light valves, which may be used in a reflective mode and therefore do not require a luminescent material, low-light situations require a backlight generated by a luminescent material. Many of the most versatile, efficient activators are rare-earth (RE) elements, for reasons that will be discussed. The ability of RE ions to emit red, green, and blue light make them well suited for application in visible-display technologies. This article reviews dopant and host material systems, excitation mechanisms, and the factors that limit the achievable luminescent intensity and efficiency. Device configurations for modern displays are discussed, as are materials and structures for next-generation technologies. Since each display technology has different performance and operational requirements, only the basic characteristics will be discussed here to enable an appreciation of emission from RE activators. References to the literature are supplied to further direct the reader to more in-depth discussions.