1 Introduction

In this paper, we describe many of the dominant and also unusual physical scientific and technical concepts enabling the series of very high-power lasers to have been built at the Livermore Laboratory (LLNL1964 to present). Footnote 1 Up to 2 MJ of UV output from the multi-beam NIF laser is being used for irradiating mm- to cm-size materials with up to 4 × 1014 W of UV light. Because of the special interest in fusion research [1, 2, 3] and the cost of the system, the NIF laser system was designed like other similarly complex scientific-user systems such as large accelerators, telescopes, gravity detectors.

Those of us who had the good fortune to be at Stanford when A.L. Schawlow and Theodor Haensch (Fig. 1) were developing modern laser spectroscopic techniques and when the Siegman, Harris, Byer groups were creating new quantum electronics devices, were very fortunate indeed. Those influences continue to this day.

Fig. 1
figure 1

Ted Haensch and Arthur Schawlow 1970

Full size image

Shortly after the laser’s description [4] and the first lasing demonstration in 1960 [5] at the Hughes Laboratory, researchers around the world began experimental and theoretical studies to understand the breath of this new optical creation. At Stanford, A.L. Schawlow and his group began investigating the laser’s uses for spectroscopy, first using accidental overlaps of laser emission lines with atomic absorption lines. Later, they developed a flashlamp-pumped dye laser for spectroscopic investigations [6]. See Fig. 2.

Fig. 2
figure 2

Holzrichter with tunable dye laser and MnF2 sample

Full size image

In the 1970,Footnote 2 Ted Haensch joined the Schawlow group and developed host of new tunable lasers and instrumentation techniques [7,8,9]. These techniques were important to early fusion laser designers by helping them to understand the spectroscopy of many materials, especially Nd3+ ions in optical crystals and glasses [10, 11].

Also in the mid 1960s, Emmett and Schawlow began conducting high-intensity experiments using the then most powerful ruby laser, to understand longitudinal and transverse Brillouin and Raman scattering from liquids and gases [12]. Their laser caused stimulated transverse optical beams to become intense enough to occasionally break their glass-holding vessels, especially if feedback of some sort from windows or edges were present. This nonlinear conversion work highlighted some of the important parasitic processes always present in very high-power laser systems, which remain a major constraint to increasing single-aperture laser power. At the same time, Siegman, Harris, Byer, and Manes [13] at Stanford, as well as many colleagues at Bell Labs and other laboratories, began studying frequency stability and communications qualities for laser communications, and also used today in our laser master oscillators. Also, other groups began investigating both pulsed and high CW power lasers for materials heating, cutting, and (surprisingly for that time) DT fusion applications (Fusion Investigators 1965).Footnote 3 Today, many prior limitations have been overcome and several large laser systems have been constructed. The largest of them, the National Ignition Facility (NIF) at the Livermore Laboratory, is being used for high-temperature materials and fusion research. It was finished in 2009 [14].

Optical coherent light generators (lasers) and corresponding optical systems have proven to be perfect instruments for communications and for delivering pin-point optical power to targeted atoms or molecules, at precise wavelengths, bandwidths, and power vs. time patterns. They have become commonplace tools for processing and assembling industrial materials, for studying materials at temperatures exceeding several million Kelvin, as well as under conditions seen only in astronomical events and nuclear explosions. Lasers have revolutionized materials cutting, drilling, welding, stress conditioning, just to mention a few applications. Exotic ideas employing lasers and coherent delivery systems for remote heating of rocket engines, mineral extraction, for transmitting space power to earth, and for nuclear fusion have been discussed. These “exotic” ideas have on the whole not yet been developed, except for laser-driven nuclear fusion [2, 3]. This topic has been under development for almost 50 years, with steady progress on this very difficult problem [15].

2 Issues regarding the generation of very high-power laser light

The physics of light propagation in vacuum permits coherent light to be generated, amplified, propagated, and focused to amazing extremes, thereby enabling enormous brightness levels on targets. However in the laboratory world, the propagation of light takes place over many meter distances through physical media such as air, glasses, reflecting and transmitting films, and crystals. Through combinations of these materials, a beam of coherent light can be created, amplified, transmitted, wavelength-converted, and focused to reach a target object.

However, these conversion processes are limited by surface and bulk materials damage along the beam path, by transverse and longitudinal reflections, and by many types of nonlinear parasitic modes. At the same time it is important to achieve efficient amplification of the laser beams, to reduce power costs, as well as to maintain low optical losses, and to use as high a fluence as possible per laser beam. This minimizes the number of beams (and costs) required to deliver a required total power to a target. To design a “best system,” experienced designers make use of past data from prior LLNL lasers. Figure 3 shows the Nova laser, prior to NIF (e.g., [16, 17]). The designers conduct detailed computer simulations, obtain specific laboratory data, and employ new technologies to minimize the costs for each laser design (they usually consider 1000 s of variations) to reach a power and energy requested by those planning future experiments.

Fig. 3
figure 3

The 100 kJ, 10-beam, IR, Green, or UV Nova Laser 1985

Full size image

One limitation of high-power pulsed lasers is that the production of high-power laser light is an inefficient process—with the total beam energy comprising only a few percent of the amplifier’s input electrical energy. This process is presently expensive, but is slowly being reduced by using arrays of semiconductor lasers instead of xenon flashlamps. Also, the propagation of laser light to a distant target (often 100 s of meters from the oscillator) is governed by optical physics such as aberrations, linear and nonlinear losses, and the use of special inventions such as relay imaging and high-power spatial filtering [18]. These also require clean space, special optics, controls, etc., which add to the total costs.

For example, in present lasers the laser beam’s transverse dimensions are restricted to about 40 cm (but 80 cm long) due to amplified spontaneous fluoresce inside the Brewster’s angled Nd:glass laser amplifier plates. (This is often called ASE.) There are typically 40 such plates and many other types of glass and crystal optics. Damage of transparent optics due to inclusions, pits, and stimulated transverse scattering via Raman or Brillouin processes occurs (see Fig. 4). Today, these limits restrict single-beam NIF laser output to about 20,000 joules of 1.05-micron laser light or ~1010 W/beam, which become 10,000 joules of UV light per beam upon harmonic conversion.

Fig. 4
figure 4

Nova optical window damaged by stimulated, transverse Brillouin scattering from perpendicular edges

Full size image

3 Large laser design

Consider a very high-energy laser pulse which is contained inside a long, narrow virtual rectangular container (i.e., a photon package that contains the electromagnetic energy) The container is about 6 m in length and 1/3 m wide on its two rectangular sides. It contains ~20,000 J of infrared electromagnetic energy, and it travels at the local speed of light ~3 × 108 m/s or slower through the laser components described in this paper (see Fig. 5). For fusion, the electromagnetic photon density is low at the front of the “box” called the “foot” and rises to a very high level in the last 1 m of this moving spatial frame.

Fig. 5
figure 5

One NIF Beam line (tan color) showing a 2-pass regenerative amplifier on the left, a power amplifier, and focusing system on right

Full size image

The total ensemble of photons from 192 beams is divided into 48 beam quads of 4 beams each, with 2 such clusters of 96 beams from each side. The 192 beams contain up to 1.9 MJ of UV energy and strike the target from two sides. Each individual beam (i.e., a virtual container of photons) remains about 6 M in length enclosed by a narrowing rectangle, as it impinges on a ~1-mm-diameter surface of matter, over a period of up to 30 ns. The two large cones, each composed of 96 beam clusters, deliver a peak power density of ~1016 W/cm2 to two sides of a target in the final several nanoseconds of the irradiation.

In summary, a pulse from the master oscillator (top left of Fig. 5) is expanded, split into 192 beams, pre-amplified, and then apodized (i.e., meaning transversely shaped) to develop a square-like, flat-topped shape, while retaining its single transverse spatial mode character. It is also pre-shaped in time to correct for amplifier saturation and other effects, and its bandwidth is “spread” to 0.1 nm to avoid Brillouin scattering in glasses and in the final crystals. It is then amplified further by a sequence of Nd:glass disk amplifiers. Then, after a sequence of single-beam amplifiers, which increase the beam energy by a factor of 109, a 20,000-J beam is transferred to the optics package to the entrance to the target chamber window. There the 1.053 micron beam, 35 cm × 35 cm in transverse dimension, exits from the transfer telescope, and it traverses a vacuum window and encounters two KDP harmonic plates, which convert the output to 351 nm (3510 angstroms). The pulse for a fusion experiment has a pre-planned time shape of a initial “foot” and final “peak,” containing about 10,000 J of UV energy. Each beam then passes through a lens, then a debris shield, and is focused to a preset location at the target. Figure 5 shows a single-beam layout, and Fig. 6 shows a sketch of the entire NIF laser system and building.

Fig. 6
figure 6

NIF schematic showing blocks of amplifiers (white), whose beams meander through a labyrinth of turning mirrors surrounding the blue target chamber. The enlarged target is shown to the right

Full size image

4 Present status

The output area of NIF’s 192 beams is 24 m2, somewhat smaller than the 30 m2 defining aperture of the optical elements. The total exit area is comparable to that of other large-aperture optical systems such as a Keck I aperture at about 75 m2. However, NIF contains approximately 7648-m-scale optical elements in its beam lines whose combined area is about 2300 m2. There are also ~40 such elements in each NIF laser beam line, which exceed the number of large optics used in most telescope optical paths by ~10×. Hence, the NIF final cost which includes both the extra optics and a target system was $3B. This is about 7–10 times the cost of a Keck class telescope due to many more large optics, a building, target system, etc. Compared to other large instrumental systems, NIF is in the same cost and complexity class category as the Hubble space telescope and the LHC accelerator at CERN.

The system is complicated, mounting and aligning 7648-m-scale optics and needing about 60,000 control points to operate. Yet, the system is highly automated and easily managed for target experiments, using modern control and optical sensor technologies.

In addition, inertial fusion experiments continue to be conducted, using remarkable imaging systems, particle probes, non-imaging detectors, a 30-picosecond backlighting system (ARC) based on petawatt laser technology [19], and other diagnostics. The NIF laser is a very successful laser design and has become a very successful experimental tool. It enables high-temperature and high-density plasmas and materials experiments to be conducted on a regular basis, by users from across the USA and abroad [20, 21].