9.2.1 Pulsed Molecular Gas Lasers

Among the most widely used high-power gas lasers are those that derive their radiation from molecular transitions. Table 9.1 includes these lasers and relevant emission characteristics. Excimer lasers can be excited either via e-beam technology or, more commonly, using electrical discharges. An excimer molecule, such as XeCl, does not readily exist in nature and is created by the interaction of energetic electrons with a gas mixture that includes a buffer gas that participates in the excitation process, such as He, and the components of the lasing molecules such as Xe and Cl.

Discharge-excited excimer lasers are available in a variety of designs that provide an ample choice of performance parameters from joule-class pulse energies, at low repetition rates, to millijoule-type lasers operating at a prf of hundreds of hertz and approaching the kilohertz regime. Commercial excimer lasers offer average powers from a few watts to hundreds of watts. Pulse duration in excimer lasers is a function of the discharge configuration and the type of excimer. Most excimer lasers provide pulses in the 20–30 ns range; however, using inductance stabilized circuits, pulses as long as 180 ns in KrF and 250 ns in XeCl have been reported (Sze and Harris 1995). Excimer lasers tend to yield wide beams with multiple–transverse-mode structures. Adoption of unstable resonator techniques can be used to improve the beam quality and linewidth performance can be improved significantly using, for example, intracavity multiple-prism grating techniques (Ludewigt et al. 1987; Duarte 1991a). For a comprehensive review on excimer lasers, the reader should refer to Rhodes (1979). Sze and Harris (1995) provide an excellent discussion of tunable excimer lasers.

Nitrogen lasers are reliable, simple to build, and easy to operate. The N₂ laser emits via two electronic transitions: one is the C³Π_u − B³Π_g transition, in the ultra-violet, and the other is the $B^3{\Pi }_g-A^3{\Sigma }_g^{+}$ transition, in the near infrared. In the ultraviolet system, three vibrational transitions, corresponding to v′ = 0 →v′′ = 0, v′ = 0 →v′′ = 1, and v′ = 1 →v′′ = 0, are observed at 337.1, 357.7, and 315.9 nm, respectively (Willett 1974). Of these transitions, the one at 337.1 nm has the largest Franck–Condon factor. N₂ lasers emitting at this wavelength often deliver energies of about 10 mJ in the pulsed regime, which is approximately 10 ns wide (full width at half maximum [FWHM]) at prfs from a few hertz up to 100 Hz.

CO₂ lasers emit in the infrared mainly in the 10 μm region. These lasers can be configured into powerful sources of coherent emission in both the pulsed and the CW regime. Information about the transitions of CO₂ lasers was provided by Willett (1974). In the case of pulsed TEA CO₂ lasers, emission occurs via the P20 (00°1 - 10°0) transition corresponding to λ = 10,590 nm and their emission bandwidth prior to narrowing can be in the 3–4 GHz range (Duarte 1985a, 1985b, 1985c).

Since molecular transitions give origin to emission bands, these lasers are tunable. Application of the line-narrowing techniques described in Chapter 7 can yield tunable narrow-linewidth emission. Narrow-linewidth tunable pulsed excimer lasers are summarized in Table 9.2 and narrow-linewidth tunable pulsed CO₂ lasers in Table 9.3.

9.2.2 Pulsed Atomic Metal Vapor Lasers

Emission characteristics of atomic pulsed metal vapor lasers are given in Table 9.4. Perhaps, the most well-known member of this subgroup is the copper laser, which is also referred to as the copper vapor laser. These lasers have found numerous applications due to their ability to emit large average powers in the green at λ = 510.554 nm and, at their secondary emission wavelength, at λ = 578.21 nm. CVLs use a buffer gas, such as He or Ne, and operate at a high prf in order to attain the necessary metal vapor pressure. Excitation of the upper ² P state occurs mainly via direct electron excitation (Harstad 1983). Output parameters from individual Cu lasers have been reported to cover a wide range of values. At pulsed energies of up to several millijoules per pulse, these lasers can emit pulses in the 10–60 ns range, at prfs from 2 to 32 kHz (Webb 1991). For instance, a specific CVL can yield an average power of 100 W, at 20 mJ per pulse, and a prf of 5 kHz (Webb 1991). Integrated CVL systems have been reported to yield average powers of up to 7 kW at a prf of 26 kHz (Bass et al. 1992). Copper lasers can also be operated at low repetition rates using copper halides to attain the necessary vapor pressures at relatively low temperatures (Piper 1978; Brandt and Piper 1981).

CVLs have been reported in a variety of cavity arrangements including plane mirror resonators and various unstable resonator configurations (Webb 1991). These lasers are useful for a number of applications, including the excitation of tunable dye lasers at low and high prfs (Duarte and Piper 1982, 1984; Webb 1991).

9.2.3 CW Gas Lasers

Perhaps the most well-known gas laser is the helium–neon (He–Ne) laser. Its most well-known line corresponds to the 3s₂ − 2p₄ transition at λ = 632.82 nm. This transition is made possible by excitation transfer from atoms at the helium metastable level He^* (2³ S₁) to ground-state Ne atoms (Willett 1974). The emission from this laser is characterized by a most beautiful TEM₀₀ beam at a laser linewidth typically less than a few gigahertz in the absence of frequency-selective intracavity optics. At the referred wavelength, and depending on the discharge length, available powers vary from a few milliwatts to a few tens of milliwatts. Additional transitions available from the He–Ne laser are listed in Table 9.5. He–Ne lasers incorporating broadband mirrors and tuning optics can emit at several visible transitions.

Another CW gas laser that is widely used in the laboratory, due to its powerful blue-green transitions, is the Ar⁺ laser. All the transitions listed in Table 9.6 are excited via electron impact (Willett 1974). The dominant CW transitions are those at 487.99 and 514.53 nm. It should be noted that relatively compact Ar⁺ lasers can be engineered to deliver powers at the tens of milliwatts range, whereas large systems have been configured to yield as much as 175 W (Anliker et al. 1977). An additional feature of laboratory Ar⁺ lasers is the option to tune from line to line using intracavity optics.

Additional CW metal vapor lasers are the He–Zn and He–Cd lasers. Ion transitions in these lasers are excited via Penning and Duffendack reactions (Piper and Gill 1975). In the first reaction, the helium metastable He^* (3S₁) interacts with the metal atoms, whereas in the second reaction, it is helium ion He⁺ (²S_1/2) that participates in the excitation. In the case of the He–Zn laser, the transitions at 491.16, 492.40, and 758.85 nm are excited via Duffendack reactions, which can only occur in hollow-cathode discharges that give origin to energetic electrons. Using a hollow-cathode He–CdI₂ discharge, Piper (1976) combined simultaneously 4 transitions from Cd⁺ and 11 transitions from I⁺ to produce a most striking TEM₀₀ white light laser beam. All of the I⁺ transitions listed in Table 9.5 participated in the emission plus the Cd⁺ transitions at 441.56, 533.75, 537.80, and 806.70 nm. All of these transitions are excited via Duffendack reactions except the 441.56 nm transition from Cd⁺ that results from Penning ionization.

For a survey of CW CO₂ lasers, the reader should refer to Willett (1974) and Freed (1995). An excellent review on frequency selectivity in CO₂ lasers was provided by Tratt et al. (1985).

9.3.1 Pulsed Organic Dye Lasers

Pulsed dye lasers are divided into two subclasses: laser-pumped pulsed dye lasers and flashlamp-pumped pulsed dye lasers.

As apparent from Table 9.7, laser-pumped dye lasers can provide either hundreds of joules per pulse at low repetition rates or very high average powers, well into the kilowatt regime at prfs in excess of 10 kHz. Excellent reviews on CVL-pumped dye lasers are given by Webb (1991) and on excimer laser-pumped dye lasers by Tallman and Tennant (1991).

The performance of laboratory-size laser-pumped dye lasers can be illustrated, considering the work of Bos (1981) who reported a linewidth of Δν ≈ 320 MHz at λ = 590 nm with a telescopic oscillator incorporating an intracavity etalon. Using three stages of amplification, the output energy was 165 mJ at an overall conversion efficiency of 55% for excitation at 532 nm. Employing a HMPGI grating oscillator and two stages of amplification, Dupre (1987) reported 3.5 mJ, and Δν = 1.2 GHz, at λ = 440 nm. The conversion energy efficiency waŝ9% for excitation at 355 nm.

CVL excitation of HMPGI grating oscillators yielded Δν ≈ 600 MHz, for λ = 575 nm, and a pulse length of 12 ns (FWHM) at a conversion efficiency of ~5% (Duarte and Piper 1984). The average output power was 80 mW at a prf of 8 kHz. Using the same class of multiple-prism oscillator, and one amplifier stage, Singh et al. (1994) reported Δν ≈ 1.5 GHz and a conversion efficiency of 40% at a prf of 6.5 kHz. Emission characteristics of narrow-linewidth tunable dye laser oscillators are summarized in Table 9.8.

The performance of large flashlamp-pumped dye lasers is summarized in Table 9.9. For a comprehensive and authoritative review of flashlamp-pumped dye lasers, the reader should consult Everett (1991). In general, these lasers have been used to generate large energies in pulses up to the microsecond regime. The energy per pulse is so high that fairly modest prfs can generate average powers in the kilowatt regime. In this regard, organic dye lasers continue to provide the best alternative for high-energy pulse generation tunable directly in the visible spectrum (Duarte 2012), which is an attractive avenue for directed energy applications.

Besides their intrinsic ability to generate large pulsed energies, flashlamp-pumped dye lasers have also been configured in small-scale laboratory versions designed to yield narrow-linewidth, highly stable, tunable laser emission. For a review on this subject, see Duarte (1995b).

A ruggedized MPL grating coaxial flashlamp-pumped dye laser oscillator yielding a single-transverse-mode laser beam with Δθ ≈ 0.5 mrad, and a laser linewidth of Δν ≈ 300 MHz, at λ ≈ 590 nm, was reported by Duarte et al. (1991). The emission was extremely pure with ASE levels a few parts in 10⁻⁷. This dispersive oscillator provided ~3 mJ per pulse, at a pulse length of Δt ≈ 50 ns, using rhodamine 590 at a concentration of 0.01 mM. Its wavelength stability was measured to be (δλ/λ) ≈ 5 × 10^-7. This organic tunable laser was engineered in an all-Invar structure and its gain medium was flowed in a hermetically sealed stainless steel and polytetrafluoroethylene system. This ruggedized oscillator (also included in Table 9.8) was shown to maintain its wavelength and linewidth characteristics following displacement on a rugged terrain (Duarte et al. 1991).

The output from narrow-linewidth oscillators can be amplified using single-stage amplifiers to yield hundreds of millijoules per pulse. Flamant and Maillard (1984) used a two-etalon oscillator to excite a flat-mirror amplifier to attain Δν = 346 MHz and a pulse energy of 300 mJ at λ = 590 nm. Using a multiple-prism grating oscillator and a single-stage unstable-resonator amplifier, Duarte and Conrad (1987) achieved Δν = 375 MHz and a pulse energy of 600 mJ at λ = 590 nm.

9.3.2 CW Organic Dye Lasers

Dye lasers have had a significant impact in high resolution spectroscopy, and other iconic applications including laser cooling, given their tunability, excellent TEM₀₀ beam quality, and intrinsic narrow linewidths that can readily reach the few megahertz level. CW dye lasers typically use Ar ⁺ and Kr ⁺ as excitation sources, although in principle they could use any compatible laser yielding TEM₀₀ emission. It should be noted that CW dye lasers have been excited with a variety of lasers including diode lasers (see, e.g., Scheps 1993). Table 9.12 summarizes the performance of relatively high-power CW dye lasers, some of which yield SLM oscillation at linewidths in the megahertz regime. Stabilization techniques resulting in the demonstration of laser linewidths as low as 100 Hz (Drever et al. 1983) are discussed in an excellent review by Hollberg (1990).

A further towering contribution of the CW dye lasers to the field of optics and lasers was their use as principal tools in the development of ultrashort pulse lasers that gave origin to the femtosecond lasers. Among the important concepts developed in this endeavor were the generation of bandwidth-limited ultrashort pulses (Ruddock and Bradley 1976), the colliding-pulse mode locking technique (Fork et al. 1981), and prismatic pulse compression (Dietel et al. 1983; Fork et al. 1984) (see Chapter 4). Using an extracavity pulse compressor, consisting of a four-grating array and a four-prism array, Fork et al. (1987) compressed a 50 fs pulse further to just 6 fs. Diels (1990) and Diels and Rudolph (2006) provided comprehensive reviews on this subject. Table 9.13 summarizes the performance of femtosecond pulse organic lasers.

9.4.1 Ionic Solid-State Lasers

Optically pumped ionic solid-state lasers include the well-known Nd laser that can exist either in a crystalline or in a glass host. These lasers are very well suited to be configured in various cavity arrangements, including unstable resonator arrangements, which yield single-transverse-mode emission. The linewidth of a TEM₀₀ laser at 1064 nm is typically 15–30 GHz (Chesler and Geusic 1972). Frequency doubling using nonlinear crystals, intracavity or extracavity, yields efficient conversion into the visible. Originally, these lasers were excited using flashlamp pumping; however, diode laser pumping has become rather pervasive. Commercially available diode laser-pumped Nd:YAG lasers can yield tens of watts at prfs in the kilohertz regime. Individual laser pulse lengths can be in the 10–15 ns range. Ionic gain media, in crystalline hosts, has also lased in the CW regime.

Nd:glass lasers are mainly operated at very high peak powers, and low prfs, in the TW regime. Some of the most well-known ionic solid-state lasers, with their respective transitions, are listed in Table 9.14.

9.4.2 Transition Metal Solid-State Lasers

Transition metal solid-state lasers include the alexandrite and Ti:sapphire lasers that are widely tunable. This quality has made the Ti:sapphire lasers particularly applicable to the generation of ultrashort pulses in the femtosecond regime. Although the ruby laser is mainly operated in the pulsed regime, typically delivering a few joules of energy per pulse, the other two media are very versatile and are well suited to both pulsed and CW operations (Walling and Peterson 1980; Walling et al. 1980a; Moulton 1986). Tuning ranges are listed in Table 9.15. A comprehensive review of transition metal solid-state lasers was given by Barnes (1995a).

All-solid-state Ti:sapphire lasers are available commercially with TEM₀₀ beam profiles, and emission in the SLM domain, delivering average powers in the watt regime at ~10 kHz. Using CVL pumping, narrow-linewidth emission has been demonstrated at average powers of 5 W at 6.2 kHz, at a conversion efficiency of ~26% (Coutts et al. 1998). Liquid nitrogen cooling of Ti:sapphire gain media has resulted in CW output powers of up to 43 W at an efficiency of ~42% for broadband lasing (Erbert et al. 1991). In the ultrashort pulse regime, Ti:sapphire lasers have been shown to deliver pulses as short as 5 fs (Ell et al. 2001).

Titanium sapphire lasers emitting in the ultrafast, or ultrashort pulse, regime are included in Table 9.16.

9.4.3 Color Center Lasers

Laser-excited color center lasers span the electromagnetic spectrum mainly in the near infrared at the 0.82 ≤ λ ≤ 3.3 μm region. Lasing has been reported mainly in the CW regime at power levels in the 10 mW–2.7 W range. An excellent review article on this subject was written by Mollenauer (1985).

Using the center F₂ in a LiF host, lasing in the 0.82–1.05 μm region was reported with a maximum power of 1.8 W (Mollenauer and Bloom 1979). Using the same center in a KF host, the same authors reported laser emission in the 1.22 ≤ λ ≤ 1.50 μm region at a maximum power of 2.7 W for an excitation power of ~5 W. Narrow-linewidth emission was reported by German (1981), using the FA(II) center in a RbCl:Li host, in a grazing-incidence cavity configuration. SLM oscillation was reported at power levels in excess of 10 mW in the 2.7 ≤ λ ≤ 3.1 μm region.

9.4.4 Diode Laser-Pumped Fiber Lasers

An area of the solid-state laser field that has evolved rapidly over the past few years is the fiber laser subfield. Glass or silica fibers doped with elements such as Yb, Nd, Er, or Tm are the active media that are increasingly excited with high-power diode lasers. From the versatility of doping elements, it should be apparent that these lasers cover the near infrared from beyond 1 μm to about 3 μm. Since the breadth and scope of this subfield is enormous, attention is only given to some of the important characteristics of fiber lasers. First, these lasers can be extremely efficient. Second, as indicated in Table 9.17, they can achieve high output CW powers under diode laser excitation. The third important quality of diode laser-pumped fiber lasers is broad continuous tunability. These qualities are well illustrated by fiber lasers included in Table 9.18.

In regard to the narrow-linewidth tunable laser reported by Auerbach et al. (2002), it should be mentioned that the authors reported extremely low levels of ASE. Also, the hybrid telescope grazing-incidence (HTGI) grating configuration (Yodh et al. 1984; Smith and DiMauro 1987) used to induce narrow-linewidth oscillation is a variant of the principle of the HMPGI grating configurations (Duarte and Piper 1981, 1984) that is described in Chapter 7. Tunable fiber lasers were reviewed in detail by Shay and Duarte (2009).

9.4.5 Optical Parametric Oscillators

Although the optical parametric oscillator (OPO) does not involve the process of population inversion in its excitation mechanism, it is included, nevertheless, since it is a source of spatially, and spectrally, coherent emission which is inherently tunable. For detailed review articles on this subject, the reader is referred to Barnes (1995b) and Orr et al. (1995, 2009). The nonlinear optics aspects of OPOs are examined in Chapter 8. Spectral characteristics of several well-known OPOs are given in Table 9.19.

OPOs are optically pumped devices that make wide use of Nd:YAG lasers and the different harmonics that can be provided with these excitation sources. A fairly instructive case study on the performance of pulsed OPOs was provided by Schröder et al. (1994) who reported an output energy of 77 mJ for an excitation energy of 170 mJ, at 355 nm, for a 15 mm-long LiB₃O₅ crystal. The emission bandwidth was ~0.5 nm (~366 GHz) at λ ≈ 640 nm. This performance was obtained using excitation pulses 5–6 ns in duration at a prf of 10 Hz. Pulsed OPOs for analytical applications tend to operate at prfs of ~10 Hz (He and Orr 2001). Barnes and Williams-Byrd (1995) discuss thermally induced phase mismatch and thermally induced lensing that tend to limit the average power in OPOs and optical parametric amplifiers.

Certainly, all the linewidth narrowing techniques described in Chapter 7 are applicable to OPOs and further details are given in Chapter 8. For example, a multiple-prism grating cavity, incorporating an intracavity etalon, in a LiNbO₃ OPO yielded a linewidth of Δν ≈ 30 MHz at λ ≈ 3.4 μm. The energy conversion efficiency was reported at ~0.74%.

Initially, OPOs in the CW regime were longitudinally excited by CW frequency-doubled Nd:YAG lasers (Smith et al. 1968). An improved approach consists in using diode laser excitation in intracavity configurations. Using fiber-coupled excitation from a diode laser, at 940 nm, Jensen et al. (2002) excited a Yb:YAG crystal whose emission was focused on a periodically poled LiNbO₃ crystal. For a diode laser power of 13.5 W, tunable emission in the 3.82 ≤ λ ≤ 4.57 μm range was reported at output powers of 200 mW (Jensen et al. 2002).

Femtosecond OPOs in linear and ring cavities, incorporating multiple-prism pulse compressors, were discussed by Powers et al. (1993). An authoritative and detailed review on the performance and output characteristics of OPOs for spectroscopic applications was given by Orr et al. (2009).

9.5.1 Tunable Quantum Cascade Lasers

A relatively new class of tunable lasers are the quantum cascade lasers often simply designated by the acronym QCLs. The principle of operation of the QCLs is outlined in Chapter 1. These lasers have had a significant impact in a broad spectral region as they have introduced a compact alternative, with all the advantages of the semiconductor technology, to produce broadly tunable coherent radiation in the mid and deep infrared. As indicated in Table 9.25, external cavity tunable QCLs have been shown to operate in the 7 ≤ λ ≤ 12 μm region.

9.5.2 Tunable Quantum Dot Lasers

Quantum dot lasers engineered with external cavities have demonstrated tuning ranges within the 1000 ≤ λ ≤ 1300 nm region as indicated in Table 9.26. A tunable quantum dot laser, using InAs as gain medium and employing a diffraction grating deployed in Littrow configuration, is reported to yield a laser linewidth of Δν ≈ 200 kHz and a tuning range of 1125 ≤ λ ≤ 1288 nm. Frequency stabilization reduced the linewidth to Δν ≈ 30 kHz (Nevsky et al. 2008).