J. J. Rocca - Table-Top Soft X-Ray Lasers

REVIEW OF SCIENTIFIC INSTRUMENTS
VOLUME 70, NUMBER 10
OCTOBER 1999
REVIEW ARTICLE Table-top soft x-ray lasers

J. J. Rocca
Electrical and Computer Engineering Department, Colorado State University, Fort Collins, Colorado 80523
Received 29 October 1998; accepted for publication 19 January 1999 This article reviews the progress in the development of practical table-top sources of soft x-ray laser radiation. The eld is rapidly approaching the stage at which soft x-ray lasers sufciently compact to t onto a normal optical table will be routinely utilized in science and technology. This is the result of recent advances in the amplication of soft x-ray radiation in both compact laser-pumped and discharge-pumped devices. The use of excitation mechanisms that take full advantage of new ultrafast high power optical laser drivers and multiple pulse excitation schemes has resulted in the demonstration of saturated soft x-ray amplication at wavelengths as short as 14 nm using several Joule of laser-pump energy. Moreover, several schemes have demonstrated signicant gain with only a fraction of a Joule of laser-pump energy. In addition, the demonstration of saturated table-top soft x-ray lasers pumped by very compact capillary discharges has shattered the notion that discharge-created plasmas are insufciently uniform to allow for soft x-ray amplication, opening a route for the development of efcient, high average power soft x-ray lasers. Recently, a table-top capillary discharge laser operating at 46.9 nm has produced millijoule-level laser pulses at a repetition rate of several Hz, with a corresponding spatially coherent average power per unit bandwidth comparable to that of a beam line at a third generation synchrotron facility. This review summarizes fundamental and technical aspects of table-top soft x-ray lasers based on the generation of population inversions in plasmas, and discusses the present status of development of specic laser systems. 1999 American Institute of Physics. S0034-6748 99 00110-0
I. INTRODUCTION
Great progress has been achieved in the last several years in the development of very compact and practical soft x-ray lasers.14 The eld is rapidly approaching the stage at which soft x-ray lasers sufciently compact to t onto a normal optical table frequently described as table-top lasers will be routinely utilized in science and technology. The motivations for the development of table-top soft x-ray lasers are numerous. They include the characterization and processing of materials, very high resolution metrology, studies in atomic physics, photochemistry and photophysics, biological imaging, and the diagnostics of very high density plasmas. Practical table-top x-ray lasers will also open up exploration of completely new elds, such as nonlinear optics at ultrashort wavelengths. In the not very distant future, when their wavelength will drop below 1 nm, applications such as time-resolved x-ray diffractometry of biological and inorganic materials, and medical diagnostics with coherent x rays will also be realized. Moreover, several of the most important applications may yet be unforecasted. When novel practical sources of intense electromagnetic radiation were developed in the past, new regimes of physical parameters became accessible, often resulting in the observation of unexpected phenomena that led to important scientic and technologic breakthroughs. Based on such historical considerations, many researchers believe that the benets of the widespread
0034-6748/99/70(10)/3799/29/$15.00 3799
availability of practical and affordable table-top x-ray lasers will impact numerous scientic disciplines and industries, even exceeding current predictions. This review summarizes fundamental and technical aspects of the development of table-top soft x-ray lasers based on the generation of population inversions in plasmas. However, the direct amplication of radiation in plasmas is not the only means by which coherent soft x-ray radiation can be generated. Alternative methods include harmonic upconversion of high power optical lasers,58 synchrotron sources,911 and free electron lasers FELs .12,13 Synchrotron sources have the very important advantages of broad tunability and high average power. In contrast, they fall short of the high peak brightness needed in applications such as the study of nonlinear phenomena at ultrashort wavelengths14 and the diagnostics of dense plasmas.15,16 Also, they are very large and expensive to construct. An alternative also based on accelerator technology is the use of self-amplied spontaneous emission in a FEL.12,13 In a soft x-ray FEL, an electron beam would radiate at higher powers and with better coherence than it does due to spontaneous synchrotron radiation. Several single-pass ultrashort wavelength FELs have been proposed or are under construction.12 Nevertheless, these lasers will not be table-top devices in the near future. Alternatively, soft x-ray coherent radiation can also be generated using either nonlinear optical techniques for frequency upconversion of optical laser radiation, or by direct amplica 1999 American Institute of Physics
Downloaded 06 Jul 2006 to 129.82.233.53. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
3800
Rev. Sci. Instrum., Vol. 70, No. 10, October 1999
J. J. Rocca
tion of spontaneous emission in a plasma. Very important progress has been achieved in the past several years in the generation of high order harmonics of intense ultrashort pulse optical lasers. Harmonic up-conversion in table-top laser systems has been reported to generate radiation at wavelengths as short as 2.7 nm.6 Presently, fairly optimized conditions in nonphase matched congurations yield typically a conversion efciency of about 10 6 in the range of 1040 eV on the order of 109 photons per pulse and 10 8 in the 40150 eV range (106 107 photons per pulse .5 The highest energy reported for a high order harmonic pulse, 60 nJ at about 50 eV, was obtained using the second harmonic of a powerful glass laser.8 Recently, the demonstration of phasematched harmonic conversion of visible light into soft x rays with a generation efciency of 10 5 10 6 in the 4070 eV spectral region7 was reported. Soft x-ray pulses with an energy of 0.2 nJ/pulse per harmonic order were produced at a repetition frequency of 1 kHz utilizing 20 fs optical pulses.7 This recent development promises to result in soft x-ray pulses with a high peak brightness and high average power. Alternatively, table-top x-ray lasers have the advantage of a much higher energy per pulse, and the potential to produce x-ray beams with very high average power.17 For example, at a photon energy of 26.5 eV, a very compact discharge pumped table-top soft x-ray laser has already produced output pulses with an average energy of 0.88 mJ ( 2 1014 photons/pulse) at a repetition rate of 4 Hz, corresponding to an average output power of 3.5 mW.18 Nevertheless, as different applications will require soft x-ray radiation with specic characteristics, in many cases the use of these various sources will be complementary. The focus of this article is table-top soft x-ray lasers. For a description of the characteristics and status of alternative methods for the generation of coherent soft x-ray radiation, the reader is referred to the literature.513 The quest for practical x-ray lasers started shortly after the demonstration of the rst lasers in 1960.19 Proposals of excitation schemes for x-ray lasers date back to 1965, when the possibility of achieving soft x-ray amplication by collisional recombination was rst suggested by Gudzenko and Shelepin.20 This was followed by proposals of photoionization pumping of x-ray lasers in 196721 and of electron impact excitation schemes.2227 The latter were inspired in part by the earlier success in the development of visible and ultraviolet ion lasers excited by electron collisions.28,29 However, the dramatic scaling of the pump power requirements with decreasing wavelength and the low normal incidence reectivity of bulk materials at soft x-ray wavelengths, combined with the short lifetime of the excited levels involved in the lasing process, made the realization of soft x-ray lasers a very challenging task. Several experiments realized during the 1970s and early 1980s yielded the observation of population inversions and gain.3037 Nevertheless, the experimental demonstration of large amplication at soft x-ray wavelengths was not realized until 1984, when Matthews et al.38 and Suckewer et al.39 observed amplication from the generation of population inversions in plasmas by collisional electron excitation and collisional electronion recombination, respectively. These pioneering demonstrations were
soon followed by numerous successful soft x-ray amplication experiments conducted using some of the worlds most powerful lasers as pump sources.4143 Subsequent experiments achieved for the rst time soft x-ray laser operation in the saturated regime,4447 and realized proof-of-principle demonstrations in several applications. These applications include microscopy,48,49 holography,50 diagnostics of dense plasmas,15,16 and the excitation of nonlinear photoluminescence in crystals.51 However, the large complexity, cost, and size of these lasers, and their low repetition rate, are barriers that limit their widespread utilization. It became clear that for soft x-ray lasers to be nearly as commonly used as optical lasers, the development of compact, low cost table-top soft x-ray ampliers was required. Several articles have reviewed the progress in the development of x-ray lasers,5257 starting with Waynant and Elton in 1976.52 The book by Elton discusses the developments up to 1990.58 A detailed account of the progress realized since that time can be found in the proceedings of two biannual conferences that focus on the development and application of soft x-ray lasers.14,5961 The next few years are likely to witness a dramatic increase in the use of soft x-ray lasers in applications as a result of the very rapid progress recently achieved in the development of compact pump sources, and the recent demonstration of large soft x-ray amplication in table-top systems. The advances in pump sources include the development of multiterawatt table-top optical laser systems based on chirped pulse amplication6265 and fast capillary discharges capable of generating highly ionized plasma columns with very high uniformity and length-to-diameter aspect ratios approaching 1000 to 1.6670 Also contributing to the progress towards practical table-top x-ray lasers is the achievement of very important reductions in the laser pump energy required for lasing,7280 to less than 1J. 7275,80 This is in part the result of the implementation of excitation mechanisms that take full advantage of the high intensity and short pulse width of new ultrashort pulse optical laser systems. Alternatively, the observation of large soft x-ray amplication in fast capillary discharge plasmas8183 has shattered the notion that discharge created plasmas are insufciently uniform to allow for soft x-ray amplication, and has opened a new route for the development of very efcient high average power soft x-ray lasers. These recent developments, coupled to advances in soft x-ray optics84,85 create the grounds for a promising future for table-top soft x-ray laser sources.
II. REQUIREMENTS FOR SOFT X-RAY AMPLIFICATION IN PLASMAS A. Pump power requirements and scaling laws
For most efcient energy extraction from a laser media it is necessary to operate in a regime where gain saturation is achieved.86 In most optical lasers the feedback provided by the optical cavity allows us to easily reach the light intensities at which the media is saturated. In contrast, the duration of the gain in soft x-ray lasers is usually shorter than the time required for the effective use of an optical cavity. This is the result of either the rapid self-terminated nature of the population inversion in transient schemes,7478 or the difculty of
X-ray lasers
3801
where the Einstein coefcient for spontaneous emission 2 scales along an isoelectronic sequence as A 21 . Considering as an example the case of a naturally broadened tranA 21 we obtain sition with g N
2
N2
In turn, the minimum pump power density required to maintain a certain upper laser level population density N 2 scales as P N 2 A 21 hc N2
3
FIG. 1. Integrated intensity of the 46.9 nm line of Ne-like Ar as a function of plasma column length in a capillary discharge amplier. The exponential increase of the intensity ceases at l 15 cm, where the saturation intensity is reached from Rocca et al., Ref. 70 .
It results from Eqs. 3 and 4 that the power density required to obtain a certain gain coefcient in a naturally broadened transition scales as P g
5
maintaining for a sufcient period of time the stringent plasma conditions necessary for amplication in quasi-cw schemes.3847,83 Consequently, soft x-ray lasers normally operate with a single-pass or double-pass amplication of the spontaneous emission through the gain medium. In such amplied spontaneous emission ampliers the spectrally integrated intensity of the laser line increases as a function of plasma length, l, as I E g e gl 1
3/2
gle gl
1/2
where g is the small signal gain coefcient, and E is a constant proportional to the emissivity.87 Ideally, this nearly exponential amplication continues until the intensity approaches the saturation intensity. For example, in collisionally excited laser systems, gain saturation is typically reached when gl 14 20.4447,83 In the example illustrated in Fig. 1, corresponding to the amplication of the 46.9 nm line of Ne-like Ar in a discharge-created plasma, the laser line intensity increases with a gain coefcient g 0.92 cm 1 until saturation is reached at l 15 cm. 70,83 In conclusion, to achieve saturated soft x-ray amplication in a single- or double-pass amplier the gain coefcient must be, depending on the length of the plasma, 13 orders of magnitude larger than that typically encountered in most visible gas lasers.88 Moreover, the fundamental physics of the generation of amplication by stimulated emission determines a dramatic upward scaling of the power density deposition required to obtain substantial gain at ultrashort wavelengths. The small signal gain coefcient is the product of the stimulated emission cross section and the population inversion density N 23,86 N N2 N1 A 21 8 g2 , g1 2
2
N2
g2 N1 , g1
In consideration of the case of Doppler broadening, the line-broadening mechanism most frequently dominant under soft x-ray laser operating conditions leads to a different but also dramatic scaling of the power requirement, P g (kT i / ) 4 . Nevertheless, assuming for illustration purposes the natural broadening case, the operation of saturated, mirrorless lasers at 50 and 5 nm can be estimated to require a pump power density 107 and 1012 times larger than a 500 nm bluegreen laser, respectively, where it is assumed that the use of the cavity reduces the gain requirement for lasing in the visible by a factor of 100. In addition, the pump power required for lasing at soft x-ray wavelengths is often further increased by refraction losses that reduce the effective gain coefcient. Another way to illustrate the large increase in excitation required for lasing at ultrashort wavelengths is to consider the plasma conditions necessary to sustain a signicant population inversion. As an example, let us consider lasing by collisional electron impact excitation in the singly ionized Ar II and Ne-like Ar IX stages of argon at 488 and 46.9 nm, respectively. While for lasing in the blue in Ar II the necessary plasma density and temperature are typically N e 5 1013 cm 3 and T e 5 eV, 29 lasing at 46.9 nm in Nelike Ar requires N e 5 1018, T e 60 80 eV. 83 This is more than a six-order magnitude increase in the N * T e product, e which account for most of the large power density increase needed for lasing in the Ne-like ion. A further reduction of the laser wavelength to 8 nm, which can be achieved, for example, by collisional excitation of Ni-like Nd, requires an electron temperature of 800 eV and an electron density of about 3 1020 cm 3. 89 The very large pump power densities required to achieve these plasma conditions, and the often present requirement of a small optical thickness in the transverse direction,58 dene gain volumes that are small compared to those of longer wavelength plasma lasers. Typical dimensions of gain media for soft x-ray lasers range from 10 to 100 m in diameter and 0.1 to 5 cm in length. An exception are capillary discharge pumped lasers, which make use of plasma columns 300 m in diameter and up to 34 cm in length.17,18,70
3802
J. J. Rocca
recognized by Chirkov,90 and later analyzed by many others.9196 Refraction causes a loss that decreases the effective gain and, in the most severe cases limits the maximum amplication length. It also affects the direction and the spatial distribution of the amplied soft x-ray beam, increasing the beam divergence, deviating the beam, and causing depending on the symmetry of the plasma sidelobes in the beam prole44,92 or annular beam patterns.17,97 Measurements of the near-eld and far-eld intensity distributions that show the effects of refraction have been performed in several soft x-ray lasers.9799 London92 has analyzed the effect of refraction on the amplication for the case of a plasma with a density gradient in one dimension, as encountered when irradiating a foil target with a line-focused laser. Treating the density prole as parabolic, he showed that the typical distance in the direction of propagation z, in which a ray stays within the lasing medium before bending out, dened as the characteristic refraction length L r , is Lr Lx n ec . n e0 8
FIG. 2. a Schematic of ray propagating along a plasma column of dimensions 2L X , L Z having a density gradient in the x direction. Refraction deects the ray. Constant properties are assumed in the y direction, which is perpendicular to the plane of the gure. b On-axis laser ux as a function of plasma length for several values of the refraction length L r as specied, assuming a gain coefcient g 5 cm 1. For L r 0.2 the exponential growth is not mantained from R. A. London, Ref. 92 .
where n e0 is the maximum electron density. The corresponding refraction angle r can be expressed as
r
n e0 . n ec
In the one-dimensional case, refraction reduces the gain by 1/L r and determines for large plasma lengths L z L r an effective gain coefcient g eff g 1 . Lr 10
B. Refraction losses
In the most commonly utilized soft x-ray laser excitation schemes the gain increases as a function of the plasma density. However, an upper limit is set by electron collisions that at sufciently high electron density thermalize the populations of the laser levels, destroying the population inversion.58 Another limit to the maximum usable plasma density is often imposed by refraction, that bends the x-ray beam out of the gain volume as a result of the variation of the index of refraction due to electron density gradients in the transverse direction. The index of refraction in a plasma is given by 1 ne , n ec 6
London expressed the dependence of the on-axis radiation ux on plasma length as a function of a parameter, the refraction gain-length dened as G r gL r . 11
where n ec is the critical density n ec m ec 2 e2

2
Therefore, for a plasma column in which the electron density decreases as a function of distance away from an axis Fig. 2 a , the refractive index increases in that direction. In such a plasma, commonly encountered in many x-ray laser ampliers, a ray propagating from a region of lower index of refraction to one of higher index of refraction is bent away from the axis and from the region of maximum gain. The problem caused by refraction on x-ray ampliers was early
From the above expression, Eq. 10 , for the effective gain it follows that the length dependence of the laser intensity corresponds to one of two distinct cases determined by whether the refraction gain length is larger or smaller than 1. When G r 1, exponential growth of the laser power is maintained until the laser intensity reaches the saturation intensity. In contrast, when G r 1, refraction sets a limit to the exponential growth and the laser power reaches a constant value for large lengths, never achieving the saturation intensity the case of the curve corresponding to L r 0.125 in Fig. 2 b . Chilla and Rocca95 extended the analysis to the case of cylindrical geometry, which is of interest for discharge-pumped lasers and axially pumped laser-created plasmas. In this case, in which the index has a gradient in two directions, refraction introduces a loss term 1/L r for each dimension and the effective gain is g eff g 2 . Lr 12
X-ray lasers
3803
FIG. 3. Simplied Grotrian diagram of Ne-like Ge showing the laser transitions of interest and the dominant excitation and deexcitation processes responsible for the creation of population inversions between the 3p and 3s levels. The laser levels are labeled in both LS and jj notations. See C. Keane, Ref. 54.
Therefore, in this two-dimensional case, exponential growth of the intensity continues until saturation of the gain is achieved only if G r 2. Several techniques have been successfully developed to reduce the detrimental effects of refraction. They include the use of curved laser targets, the reduction of sharp density gradients by the use of a prepulse in laser-pumped systems,100117 and the use of a magnetic eld in dischargepumped lasers.118 Plasma waveguides having a density minimum on axis have also been developed74,119123 and successfully used in soft x-ray amplication experiments.74 Plasma waveguides can be expected to have an important impact on the development of efcient table-top x-ray lasers.
III. POPULATION INVERSION MECHANISMS A. Collisional electron excitation
The collisional electron excitation scheme was one of the rst explored theoretically in detail.2227 Several compact, high gain, laser-pumped7579 and dischargepumped8183,124 soft x-ray ampliers have been successfully developed based on this scheme. This x-ray laser excitation mechanism resembles that of some of the most widely utilized visible and ultraviolet ion lasers, the cw argon ion and krypton ion lasers,28,29 in which the laser upper levels are predominantly excited by direct electron impact collision from the ground state of the ion stage of interest. In the most common implementation of these lasers, the generation of a population inversion occurs in a quasi-cw regime, aided by the very favorable radiative lifetime ratio between the laser upper and lower levels. The upper levels are metastable with respect to radiative decay to the ground state, and the laser lower levels are depopulated by strong dipole-allowed transitions. The rst successful demonstration of lasing at soft x-ray wavelengths utilizing this approach was realized at Lawrence Livermore National Laboratory involved the
3 p 3s transitions in Ne-like Se and Ne-like Y.38 Ne-like ions have a fully occupied outer shell and are hard to ionize. This is a signicant advantage because it results in a high relative abundance of the lasing ions over a wide range of plasma parameters. To date, amplication has been demonstrated in the majority of the Ne-like ions having atomic number between Si Ref. 125 and Ag.126 Figure 3 shows a simplied energy level diagram for a typical Ne-like system that illustrates the laser transitions and the dominant processes involved in the generation of amplication. It corresponds to the Ge XXIII laser rst demonstrated at the Naval Research Laboratory by Lee et al.40 The 3p laser upper levels are populated by electron monopole collisional excitation E 0 from the Ne-like ion ground state and also through recombination from the F-like ion, and by cascades from higher energy levels. A small contribution might also come from inner shell ionization of the Na-like state. The population inversions are maintained by the very rapid radiative decay of the 3s laser lower levels to the ground state of the ion through strong dipole-allowed transitions E 1 . Therefore, operation of these lasers in a quasi-cw regime requires the plasma to be optically thin for the transitions originating from the laser lower level. This imposes a restriction in the maximum plasma column diameter, that is often signicantly relaxed by a Doppler shift in the lines caused by radial velocity gradients in the plasma.127129 The pumping rate and the population inversion increase with the electron density. However, as mentioned before, the maximum electron density at which an inversion can be maintained is limited by collisional mixing between the upper and lower laser levels.58 From the beginning of the study of soft x-ray lasing in Ne-like ions, models have predicted the largest gain to occur in the transition linking the (2p 5 3 p 1 S 0 ) and (2p 5 3s 1 P 1 ) upper and lower levels, respectively (J 0 1 line . This is because the highest monopole excitation rate into the 3p manifold is to the J 0 level. However, the rst observation of amplication occurred instead in the J 2 1 lines of Nelike Se, in 1984.38 Subsequent experiments conducted in the following few years conrmed the observation of larger gain in the J 2 1 lines,44 with the exception of experiments conducted in lower-Z elements that showed slightly higher gain in the J 0 1 line.40 This apparent contradiction between theory and experiment in the distribution of the gain in high Z elements attracted signicant attention, and gave origin to a number of suggested explanations.130132 Increased understanding of the cause of this anomaly resulted from recent experiments in laser-created plasmas that made use of a prepulse in the excitation. The prepulse signicantly changes the spectral composition of the x-ray laser output, resulting in a great increase in the intensity of the J 0 1 line.103108 An important characteristic of plasmas created following a prepulse is the reduced density gradients, and consequently reduced refraction. It was concluded that lasing in the J 0 1 line takes place in regions of higher plasma density, where without a prepulse refraction losses signicantly decrease the effective amplication of this line relative to the J 2 1 line. The J 0 1 line has been observed to be clearly dominant in low-Z Ne-like ions generated by
3804
J. J. Rocca
either laser excitation of a solid target using a prepulse,103,104,108 laser excitation of a gas-puff target,133 or fast capillary discharges.8183,124 The collisional electron excitation of Ne-like ions is one of the most studied and proven excitation mechanisms. It has originated some of the most robust soft x-ray lasers demonstrated to date, including several table-top soft x-ray lasers.17,76,78,83,124 Nevertheless, it also has a signicant disadvantage: the high degree of ionization and consequently the large pump power that is required to obtain lasing at a given wavelength. Alternatively, the Ni-like sequence has been proposed and successfully utilized to extend collisionally excited lasers to shorter wavelengths.43,44,77,115,116,134144 Lasers in the 3d 9 4d 3d 9 4p transitions of Ni-like ions are direct analogs to lasers in 2p 5 3p 2p 5 3s transitions in closed shell Ne-like ions, but have the advantage of producing amplication at a shorter wavelength for a given state of ionization. This higher quantum efciency signicantly reduces the pumping energy required to achieve lasing by collisional excitation at a selected wavelength. The 4d levels are populated through a combination of direct electron monopole collisional excitation from the 3d 10 ground state of the ion and from cascading from upper levels. As in the case of Ne-like lasers, the laser lower level is rapidly depopulated by dipole-allowed radiative decay to the ground state. The highest gain is observed in the J 0 1 line. Nilike soft x-ray lasers were rst demonstrated in 1987 in an Eu laser-created plasma, producing an amplication of gl 4 at 7.1 nm.43 Subsequently, the scheme was isoelectronically extrapolated to other ions with laser wavelengths as short as 3.56 nm in Ni-like Au.126 Hagelstein rst proposed the use of low Z Ni-like ions to develop table-top collisional lasers at wavelengths near 20 nm and computed signicant gain for overheated plasma conditions.138 For several years the maximum amplication obtained in Ni-like ions remained smaller than that obtained in Ne-like systems, and below the values required for gain saturation. Nevertheless, the use of multiple-laser-pulse excitation techniques and the optimization of target geometries has greatly increased the gain in Ni-like ions.141143,144 Recently, gain saturated operation has been obtained at wavelengths as short as 7.3 nm.114,115,143 The amplication of soft x-ray radiation by collisional excitation is not limited to Ne-like and Ni-like ions. Gain has also been observed in Co-like ions,140 and the use of the Nd-like sequence has also been proposed.138,145
1. Transient electron collisional excitation
picoseconds by heating the plasma at a rate faster than the relaxation rate of the excited states. The larger gain coefcients are a consequence of several phenomena that, for a short period following rapid transient heating, contribute to generate a much higher population inversion. These include the fact that immediately following rapid collisional excitation, and before collisions have the time to redistribute the populations, the upper laser level population is larger as a result of its larger rate of excitation. The other phenomena that contribute to an increased gain are the increased rate of electron excitation in an overheated plasma recall that the excitation rate depends exponentially on the electron temperature , and a reduction in the rate of collisional mixing between the closely spaced active levels. The latter allows operation at signicantly higher electron density, and therefore at increased pumping rates. Transient gains in excess of 100 cm 1 have been predicted theoretically.146150 The shortlived transient population inversions, which are pumped directly from the ground state by electron collisional excitation, last until collisions redistribute the populations among levels. Following this short transient period of high population inversion the gain decreases until the laser level populations nally reach the quasisteady state value. In summary, in the transient electron collisional excitation scheme a population inversion is created because the transient collisional processes of excitation of the laser upper and lower level occur at different rates, and not because of the faster rate of radiative relaxation of the laser lower level. Therefore, in the transient regime there is no need to limit the transverse dimension of the plasma in order to ensure optical transparency of the lower level radiation. The main advantage of the transient excitation scheme for the realization of table-top x-ray lasers is the greatly reduced laser pump energy required for excitation. For example, a few J/cm deposited in 1 ps in a line-focused plasma have produced gains of up to 35 cm 1 in the 3p 3s J 0 1 line of Ne-like Ti at 32.6 nm.76,78 The transient excitation scheme has been also extended to the Ni-like sequence, where a gain of 35 cm 1 was reported in the 4d 4p J 0 1 line of Ni-like Pd at 14.7 nm utilizing 5 J of excitation energy.77,78
B. Collisional recombination
The collisional electron excitation scheme described above is intrinsically a quasisteady state scheme in which lasing can occur for as long as the plasma conditions necessary for the generation of a population inversion can be maintained. A recently demonstrated variation of the collisional electron excitation scheme takes advantage of the much larger population inversions that can be obtained under rapid transient excitation.7678 It was rst recognized by Afanasiev and Shlyaptsev146 that gain coefcients that are 12 orders of magnitude larger than those obtained for the same transition in the quasisteady state regime can be produced for a short period of time typically subpicosecond to tens of
This population inversion mechanism was rst proposed by Gudzenko and Shelepin in 1965.20 The rst report of large amplication at soft x-ray wavelengths by plasma recombination (gl 6), corresponds to an experiment realized by Suckewer et al. in 1984.39 More recently, recombining plasmas pumped by table-top lasers has produced amplication in transitions to the ground state72,74 with gains up to gl 6.5 at wavelengths as short as 13.5 nm. Evidence of amplication by plasma recombination has also been reported in plasmas created by compact electrical discharges.66,151155 In this scheme, the laser upper level is populated following the recombination of ions with a charge Z 1 with an electron, through a three-body interaction described as collisional or three-body recombination: AZ
1
2eA Z * e.
13
X-ray lasers
3805
This type of process is the inverse of collisional electron ionization. It preferentially populates highly excited bound levels A Z * of the ion of charge Z, favoring the generation of population inversion the recombination rate scales with n, the principal quantum number of the level, as n 4 ). 156 As the reaction above suggests, the collisional recombination rate R 3br is proportional to the square of the electron density. The recombination rate is also extremely sensitive to the electron temperature R 3br T e 4.5 . 157 Therefore, the generation of large population inversion by recombination requires a dense and relatively cold plasma. Hydrogen-like ions, as originally suggested by Gudzenko and Shelepin,20 have a very favorable energy level structure for the generation of population inversions by collisional recombination. In principle several transitions can be inverted, but initially much of the attention focused in the very favorable 32 transition of these ions. As an example, Fig. 4 shows the atomic processes involved in the generation of population inversion in the 32 transition of H-like C at 18.2 nm by collisional recombination. When the plasma cools, collisional recombination populates highly excited states, and electron collisions rapidly transfers the population to levels of lower energy. Since collisional electron deexcitation is inversely proportional to the square root of the energy difference between the levels,158 this electronic cascade reaches a level at which electron de-excitation is no longer dominant over radiative decay. At this level the bottleneck that is produced in the cascade creates a population inversion respect to a lower level, that is most commonly de-excited by very rapid radiative decay to the ground state of the ion. The amplication of soft x rays by plasma recombination would seem to require conicting plasma conditions: a very highly ionized plasma and a very cold electron temperature. In practice, the problem has been traditionally solved utilizing a two-step process. First, a highly ionized and dense plasma is generated by a heating pulse. Second, the plasma is rapidly cooled following the termination of the excitation pulse. By this method the plasma is hot during the excitation pulse, allowing for the generation of the required high population density of ions A Z 1 , and cold during the recombination phase. The required cooling rate is determined by the recombination rate. The plasma can be rapidly cooled by an adiabatic expansion,30,31,37,127,159166 by electron heat conduction to a nearby wall or colder neighboring plasma,167 or by radiation from high-Z ions introduced as impurities into the plasma.168,169 All three cooling mechanisms, or combinations of them, have been utilized experimentally to generate gain at soft x-ray wavelengths by collisional electronion recombination. In particular, important efforts have been devoted to the demonstration and study of amplication in the 18.2 nm 32 line of H-like C. Initial experiments observed population inversions in plasmas that were generated by ablating solid carbon targets or carbon bers with high power laser pulses, and cooled by adiabatic expansion.30,31,3537 Large amplication was rst observed in an experiment in which a 300 J pulse from a CO2 laser with about 75 ns pulse width was used to generate a nearly totally ionized carbon plasma column by bombarding a carbon solid target im-
FIG. 4. Simplied Grotrian diagram of H-like C showing the processes responsible for the generation of population inversion between the n 3 and n 2 levels in the 18.2 nm C VI recombination laser.
mersed in a strong solenoidal magnetic eld.39 The magnetic connement was allowed to maintain a high electron density while the plasma was cooled by radiation and electron heat conduction to adjacent cooling blades. Laser pulses of about 3 mJ were generated at an efciency of 10 5 . Amplication has also been demonstrated in numerous experiments in which line focus plasmas were cooled by free adiabatic expansion.41,160165 Gain has been reported in H-like Na ions,163 in the 4 f 3d and 5 f 3d lines of Li-like ions,164166 and in Be-like and Na-like ions.170 An important advantage of the recombination scheme with respect to the collisional excitation scheme involving n 0 transitions is its more rapid scaling to shorter wavelengths with nuclear charge Z. However, recombination lasers have suffered the problem of not scaling adequately with plasma column length. The largest gain length product achieved to date is gl 8.39 It is possible that the problem is related to the very high sensitivity of the gain in recombination schemes to the variation of the plasma parameters that can be more pronounced for longer plasma columns. Nevertheless, the causes of this limitation are not well understood171 and are a topic of current discussion that must be resolved for x-ray recombination lasers to achieve their full potential.
1. Recombination lasing in transitions to the ground state
Jones and Ali172 demonstrated theoretically that large transient gains could be generated in the 21 transitions of H-like ions following recombination of a totally ionized plasma of arbitrarily low temperature. This kind of laser scheme using a transition to the ground state is very attractive to rapidly scale recombination lasers to very short wavelengths. It provides a method to amplify with a high quantum efciency very short wavelength lines in species which are only moderately ionized. For example, in this scheme lasing at 13.5 nm requires an initial plasma of totally stripped Li
3806
J. J. Rocca
ions three times ionized atoms , while lasing at 3.37 nm can be attempted in C VI in a plasma that initially is only six times ionized. In the rst case the quantum efciency QE , computed as the ratio of the laser photon energy and the energy required to totally strip the atom, is QE 0.45. This is a large improvement with respect to a recombination laser based on the 32 transition of H-like ions, that for lasing at a similar wavelength would require an inversion in H-like N ( 13.4 nm) with a QE 0.06. The high QE of laser transitions to the ground state offers a possible route for the development of very short wavelength lasers with small excitation energy, potentially resulting in very short wavelength table-top recombination lasers. However, for lasing to occur in a transition to the ground state, recombination has to be very rapid to allow for the generation of large upper laser level populations before the ground state ion level is signicantly populated. The recombination time must then be shorter than the radiative lifetime of the laser upper level which is, for example, 26 and 1.6 ps for the n 2 level of H-like Li and H-like C, respectively. These time scales are often shorter than the time required by any plasma cooling mechanism to cool a hot plasma before the ground state is signicantly populated. Therefore, amplication in transitions to the ground state requires the generation of highly stripped ions in a plasma with a low electron temperature. Peyraud and Peyraud173 proposed the use of multiphoton ionization to produce a plasma of fully stripped ions and cold electrons on a time scale that is short compared to the recombination time. However, means for the generation of initial plasma conditions for lasing in a transition to the ground state did not exist in 1975, the time of the analysis of Jones and Ali. The advent of very powerful subpicosecond lasers6265 and the understanding of above-threshold ionization174 have made possible the relatively recent observation of soft x-ray amplication in transitions to the ground state.72,74,175 Burnett and Corkum176 recognized that a cold plasma of highly stripped ions in which recombination can occur in the subpicosecond time scale can be created by optical-eldinduced OFI ionization using linearly polarized light, and proposed the realization of ultrashort wavelength lasers utilizing this approach. For a linearly polarized laser pulse, the electrons are compelled to return their quiver energy to the laser eld. This method can therefore create plasmas that are simultaneously highly ionized and cold in a time scale much shorter than their recombination time. Amendt and Eder and Eder et al.177,178 analyzed different aspects of the implementation of soft x-ray recombination lasers based on OFI. The rst observation of gain in an OFI plasma was reported by Nagata et al. in the 21 transition of H-like Li at 13.5 nm.72 Experiments subsequently conducted at Princeton179 and at Berkeley178 observed similar results. Heating associated with Raman backscattering was recognized as a possible obstacle for extending this scheme to much shorter wavelengths. For schemes requiring pump laser intensities on the order of 1017 W cm 2, e.g., Li-like Ne at 9.8 nm, it has been suggested that the use of short driving pulses 100 fs can provide a solution.178 However, the same theoretical analysis concluded that it will be difcult to keep heating associated
to Raman scattering at acceptable levels in shorter wavelength schemes that require intensities on the order of 1 1018 W cm 2. 170 Experiments are needed to evaluate the implications of these potential limitations to the scaling of OFI recombination lasers. The progress in the development of OFI recombination lasers is summarized in a later section.
C. Photoionization and resonant photopumping
These two population inversion mechanisms for x-ray amplication have in common that the excitation of the laser upper level population involves the use of high energy photons. The generation of large population inversions following the selective x-ray photoionization of inner shell electrons was originally proposed by Duguay and Rentzepis in 1967.21 The generation of population inversions by this mechanism is possible because at photon energies just above the threshold for inner shell photoionization the cross section is an order of magnitude larger for inner-shell electrons as compared to outer-shell electrons. This scheme has the potential advantage of leading to relatively compact lasers with wavelengths shorter than 1.5 nm.180,181 In principle, it can allow for operation at a low plasma temperature of less than 1 eV with consequently small Doppler broadenings and large gain coefcients. The incoherent x-ray photons that pump the laser media would be normally produced by a nearby plasma created by heating a target made of a high-Z material such as gold with an intense ultrashort laser pulse. Experiments have demonstrated total laser energy to incoherent x-ray energy conversion efciencies of 20% 1% into x rays with energy above 1 keV . Photons at energy below the inner-shell binding energy are removed with an appropriate lter to avoid pumping of the laser lower level. The rst proposal for inner-shell photoionization lasers at x-ray wavelengths21 has preceded by decades the development of sufciently powerful ultrashort pulse laser drivers.6265 The photoionization scheme was rst demonstrated in the visible region of the spectrum by Silfvast et al. in 1983.183 McGuire proposed the production of population inversions by selective Auger decay following inner-shell photoionization.184 In 1986 Kapteyn, Lee, and Falcone demonstrated lasing at 108.9 nm in doubly ionized Xe by Auger decay following the photoionization of neutral Xe.185 Output energies of several J in this transition were subsequently obtained at a 2 Hz repetition rate by Sher et al. utilizing traveling wave excitation.186 Amplication in the equivalent transition in Kr at 90.7 nm was also observed.187 However, no demonstration of a photoionization laser at x-ray wavelengths has yet been realized. Nevertheless, the recent development of relatively compact lasers capable of producing peak powers of up to 100 TW greatly increases the likelihood that this type of x-ray laser will be realized in the near future. Kapteyn has analyzed the possibility of obtaining lasing by preferentially photoionizing the K-shell electrons of low-Z elements,180 focusing in particular on the K line of Ne ( 1.5 nm). Previously, Elton had discussed the possibility of obtaining quasistationary population inversion K transitions,182 and Hagelstein had analyzed the possibility of
X-ray lasers
3807
producing a 27 eV laser by photoionizing the 2s shell of neutral Ne to create a 2p 2s inversion in Ne II.53 A main advantage of the K scheme is its scalability to very short wavelengths. The energy level diagram of the K Ne laser is illustrated in Fig. 5. The ltered x rays primarily photoionize the inner-shell electrons, producing population inversion and gain in the allowed radiative transition between the (1s) 1 2S and (2p) 1 2P levels of the singly charged ion. Rapid Auger decay with a rate of 2.7 fs 1 is the dominant mechanism of depopulation of the (1s) 1 laser upper level. However, the self-terminating lifetime is not the dominant process that limits the duration of the gain. Instead, the magnitude and duration of the gain is limited by electron collisional ionization of neutral neon atoms. The energetic electrons created in the lasing material by photoionization and Auger decay create predominantly ground-state Ne ions that are the laser lower level. Therefore, this is intrinsically an ultrashort pulse x-ray laser. Kapteyn has concluded that lasing in the K transition of Ne at 1.5 nm could be produced by traveling-wave excitation using a pump laser generating 10 J pulses of approximately 50 fs duration.180 In this scheme, the target would be a thin beryllium foil, which is coated on one side with a structured x-ray emitting material and supports a frozen mix of NeH on the other side. The beryllium acts as an absorber for low energy x rays that are capable of producing outer-shell ionization, and also separates the hot x-ray emitting region from the laser region so that thermal ionization of Ne does not take place. Hydrogen acts as an electron moderator in the lasant material to mitigate electron collisional ionization. The theoretical study of the K photoionization lasers was extended by Eder and co-workers178,181,188 who calculated a scaling law relating the necessary laser pump energy E L to the wavelength that can be achieved. Assuming 20 fs x-ray pump pulses, they obtained E L (4.5/ ) 2 , where the laser wavelength is in nm and E L is in J.189 Calculations for inner-shell photoionized lasing in C at 4.5 nm suggest that a driving laser energy on the order of 1 J should be sufcient to produce a gain coefcient of 10 cm 1 and a large gain-length product.188 Barty et al. have proposed another approach to the pumping of inner-shell x-ray lasers, in which the pumping is not based on x-ray photons but rather on a burst of fast electrons produced by an ultrashort pulse laser.63 A major motivation for this scheme is the nearly 100% efciency with which a circularly polarized high-intensity optical laser can produce a burst of electrons with a pulse width and energy roughly corresponding to the duration and ponderomotive energy of the irradiating laser, respectively. Nevertheless, atomic electron impact ionization cross sections favor outer-shell ionization over inner-shell ionization, and thus the automatic population inversion that is possible with x-ray pumping no longer takes place in this scheme. However, the scheme proposed by Barty et al. relies instead on the CosterKronic mediated decay of the laser lower level to create the inversion.63 A different pumping mechanism also based on the excitation by energetic photons involves resonant photoabsorption.190,191 In contrast with the photoionization scheme discussed above, which can make use of a broadband x-ray
FIG. 5. Simplied energy level diagram of a proposed inner-shell photoionization laser in the K transition of Ne at 1.5 nm. X rays photoionize the 1s electrons of the Ne atoms creating a population inversion and gain in the (1s) 1 (2p) 1 transition of singly ionized Ne. Rapid Auger decay depletes the (1s) 1 laser upper level. Electron impact ionization populates the (2p) 1 laser lower level limiting the duration of the gain from H. C. Kapteyn, Ref. 180 .
pump source, resonant photopumping relies on very intense line emission to pump the laser upper level. Therefore, this mechanism requires a precise wavelength coincidence between powerful pump lines and lines that can populate a laser upper level by resonant photoabsorption from a highly populated state, typically a ground state. It was recently shown by Nilsen et al. that the amplication observed in the 3d 1 P 1 3p 1 P 1 line of Ne-like Ar at 45.1 nm in a lasercreated plasma was assisted by photopumping due to reabsorption of the 3d 1 P 1 2p 1 S 0 resonance line.192 This mechanism of self-pumping of a strong emission line in an optically thick plasma has also been proposed as a mechanism for driving the population into the laser upper levels of other Ne-like and Ni-like ions.193 However, a directly photopumped x-ray laser has not yet been realized. This excitation mechanism has nevertheless produced amplication at ultraviolet wavelengths,194196 and a large number of laser proposals based on x-ray line-coincidence studies are awaiting realization.53,58,197202 Signicant efforts have been dedicated to attempt the demonstration of amplication at 23.1 nm in He-like Ne by resonant photoexcitation of the 1s 2 1s4p 1 P 1 transition in that ion with 1.1 nm radiation from He-like Na plasmas generated by very large pulsedpower generators.203205 An experiment conducted in 1992 at Sandia National Laboratory using the Saturn generator, at the time the worlds most powerful pulsed-power machine, produced 200 GW in the 1.1 nm Na pump line and obtained evidence of a population inversion in He-like Ne.205 However, no gain was reported. Of greater interest for the development of table-top lasers are experiments that have been conducted with more compact drivers.206208 Gain on the order of 1 cm 1 and a gl 2 was reported in the 3d 4 f and 3d 5 f lines of Li-like Mg ions, pumped by a Si plasma created by a laser pulse of 120 J energy and 400 ps duration.206 Another experiment observed an enhancement of the population of the upper level of a Mo VII transition near 60 nm when photopumped by a line in Mo XII from an adja-
3808
J. J. Rocca
cent laser-created plasma.207 Also, a resonant photopumping study of Be-like Mg IX ions by Li-like Al XI line radiation has been performed driving an Al pinch with a 0.5 TW pulsed power generator.208 The objective of this experiment was to obtain amplication on the Mg IX 2s4 p 2s3d transition at 2.28 nm. Evidence of uorescence from the 4p level of Mg IX due to photopumping by the 4.83 nm line of Al XI was reported. The population inversion mechanisms discussed above are by far the most studied, and include all the soft x-ray lasers demonstrated to date. However, other x-ray population inversion schemes based on different atomic processes such as charge transfer, have also been proposed.58,209211 Recently, Hagelstein analyzed the possibility of extending the concept of lasing without inversion to soft x-ray wavelengths.212
IV. PRESENT STATUS OF TABLE-TOP DRIVERS
Many of the recent advances in table-top soft x-ray lasers have been made possible by the development of new compact high power drives. Up to the late 1980s practically all soft x-ray amplication experiments made use of highenergy optical laser drivers that generated pulses of nanosecond duration. Large pulsed-power generators were also used in attempts to obtain soft x-ray amplication, but did not succeed in observing gain. While research towards the development of table-top x-ray lasers based on nanosecond pulse laser drivers continues to date,80 much of the recent effort has shifted towards the use of small-scale high power ultrafast laser drivers7478,216 and compact high power discharges.17,6870,8183,124,152155 New opportunities for the efcient pumping of table-top soft x-ray lasers have arisen from the development of compact, very high power ultrafast optical lasers based on the technique of chirped-pulse amplication CPA , and from the demonstration of highly ionized capillary discharge plasma columns of unprecedented axial uniformity. CPA of ultrashort laser pulses presently allows for the generation of multiterawatt peak powers from tabletop systems operating at repetition rates of typically 10 Hz.6265 Compact fast capillary discharges have been shown to successfully concentrate large amounts of electrically stored energy in needle-shaped plasma columns with length to diameter ratios approaching 1000:1 and excellent axial uniformity.6870,8183 The characteristics of these drivers for table-top lasers are summarized in the following sections.
A. High power ultrashort-pulse lasers
The high power ultrashort pulse drivers used in table-top x-ray laser research are based on either excimer-based or solid-state amplier media. Ultrashort pulse excimer based systems have been available since the 1980s and typically produce pulses of 0.30.5 ps duration and energy up to a few hundred mJ.213215 However, the output energy of relatively compact high power excimer ampliers is limited to these values by the low saturation uence mJ/cm2 . Presently, the majority of compact high-peak power laser systems use CPA in solid-state ampliers, a scheme rst demonstrated by Mourou and co-workers217 based on the concepts of optical
pulse compression developed by Treacy218 and Martinez.219 This approach allows us to take full advantage of the high saturation uences 1 J cm2 , broad gain bandwidth, and relatively long upper laser level lifetime of solid-state amplier media such as Ti:sapphire, Ndglass, and Cr:LiSAF. The result has been the development of terawatt peak-power optical lasers operating at pulse repetition rates of 10 Hz that can t on a single optical table, and larger systems producing several tens of terawatts that t in a few optical tables.6265 The characteristics and design issues of high power ultrafast laser systems have been recently reviewed in detail by Backus et al. in a previous issue of this journal.62 Nevertheless, for completeness the basic concept and status of CPA high power ultrashort-pulse laser drivers for table-top soft x-ray laser research are briey summarized below. In the CPA technique, a short seed pulse is rst stretched in duration, then amplied, and nally recompressed to ideally its initial duration.217 The process allows us to obtain extremely high peak powers by the amplication of ultrashort pulses, while avoiding very high intensities in the amplication pulse that would damage the ampliers.62,64,217 A schematic description of a CPA amplier system is illustrated in Fig. 6. Low-energy seed pulses (1 10 10 9 J) are typically produced by a mode-locked oscillator, most commonly a Ti:sapphire laser pumped by a cw Ar ion laser or a diode-pumped frequency-doubled Nd:YAG laser. The discovery of Kerr lens mode locking220,221 and its subsequent renements,222 have produced pulses as short as 4 fs and have made the generation of sub-20 fs seed pulses a routine. Prior to amplication, the seed pulse is stretched in duration by a factor 103 104 by introducing a frequency chirp, with a corresponding reduction in intensity. The concepts of dispersion control developed by Martinez219 allow for the construction of a perfectly matched stretchercompressor pair. The stretcher, that in its original design consisted of a pair of diffraction gratings in an antiparallel conguration containing a telescope composed by two identical lenses, generates large positive dispersion factors. For the amplication of very short seed pulses it is necessary to implement a stretcher that matches the dispersion of the compressor and of all other dispersive elements in the system, including the gain medium, Pockel cells, polarizers, isolators, etc. This has been accomplished with all-reective stretchers, which present several advantages.62,223 The stretched seed pulses are rst amplied to 110 mJ by either a regenerative or a multipass amplier, and are subsequently injected into one or more ampliers. The design of ultrashort-pulse laser ampliers must take into account the gain narrowing that results from the reduction of the bandwidth of the amplied pulse caused by the larger amplications of the frequency components, which are closer to the central frequency of the gain medium.62 However, gain narrowing has not precluded the amplication of sub-20 fs pulses to peak powers up to 100 TW.62,65,224 Following amplication, the pulse is nally recompressed close to its original pulse width by a pair of identical parallel gratings that provide a negative group delay dispersion. In very high power systems, at least the nal grating compressor is enclosed in a vacuum chamber to avoid wave front distortion and lamentation.
X-ray lasers
3809
FIG. 6. Schematic diagram of a high-power ultrashort-pulse laser system based on chirped pulse amplication. The shaded rectangles in the expansor and compressor setups represent diffraction gratings from Backus et al., Ref. 62 .
Relatively compact ultrashort pulse optical laser systems with peak powers in the 5100 TW range have been recently demonstrated based on the CPA technique.6265 A schematic of a multi-TW laser system developed at the Max Born Institute is illustrated in Fig. 7. It consists of a hybrid CPA system utilizing a Ti:sapphire oscillator and regenerative amplier and a Nd:glass amplier chain designed to produce a synchronized sequence of a long pulse 1.2 ns of up to 7 J and a short pulse 0.7 ps of up to 4 J .76,225 In this system, a seed pulse produced by a Ti:sapphire oscillator is chirped to 1 ns in a double pass grating stretcher. A single pulse is switched out of the oscillator and is amplied in a regenerative amplier. The output pulse is further amplied in a chain of ve Nd:phosphate glass ampliers before being split to form two beams that are amplied in two nal amplier arms. The pulse from one arm is compressed in a grating compressor that is enclosed in an evacuated box to produce a short pulse of 0.7 ps duration. The two pulses have orthogo-
nal polarization in order to recombine them with a polarizer prior to focusing them onto the solid target. This implementation allows a jitter-free synchronization of the two pulses. As discussed later, this type of laser system has been successfully used to generate very large transient population inversions in Ne-like and Ni-like ions.7678,225,226 Great progress has also been recently achieved in the development of sub-20 fs systems that are likely to lead to the demonstration of gain in inner-shell transitions in the near future. A 50 TW, sub-20 fs Ti:sapphire system capable of operating at 10 Hz has been demonstrated at the University of California at San Diego63 and a 100 TW, 19 fs, 10 Hz system has been recently completed at JAERI Japan .224 Systems with similar characteristics and peak powers up to 30 TW have also been built at several other institutions.64 For more details on high power ultrafast lasers, the reader can refer to the review paper by Backus and references therein.62 While some of the above lasers represent the state of the art in terms of output powers for relatively compact systems, smaller ultrashortpulse lasers with lower output powers have also made signicant contributions towards the development of table-top soft x-ray lasers,7275 and are likely to continue playing a role in the future.
B. Fast capillary discharges
Direct excitation of plasma columns with an electrical discharge has the potential advantage of generating soft x-ray lasers that are very efcient and compact. However, axial inhomogeneities in the plasmas of high power dis-
FIG. 7. Schematic diagram of the high power CPM hybrid laser system developed at the Max Born Institute to pump transient inversion soft x-ray lasers by a sequence of a long and a short pulse. This laser was successfully used to pump transient-inversion collisional lasers in Ne-like Ti and Ne-like V Ref. 76 . Diagram courtesy of P. V. Nickles, MBI .
3810
J. J. Rocca
charges, produced by nonsymmetric compressions and instabilities, have impeded for many years the development of discharge driven soft x-ray lasers. Recently, this situation has dramatically changed with the demonstration of saturated table-top lasers driven by capillary discharges.83 Fast discharge excitation of capillary channels 1.54 mm in diameter lled with preionized material has resulted in the successful compression of a signicant fraction of the discharge power into plasma columns of 300 m diameter, achieving the high power density deposition and the good axial uniformity necessary for soft x-ray lasing.68,69,82 Fast compression of larger diameter plasmas in Z-pinch discharges has also generated axially uniform columns227 in which soft x-ray amplication by collisional recombination has been investigated.153 Rocca et al. rst proposed the development of soft x-ray lasers based on fast capillary discharges in 1988.228 Prior to their use in soft x-ray laser research, discharges in evacuated plastic capillaries had been investigated for applications in spectroscopy, microscopy, and lithography.229232 However, with the exception of an experiment in an evacuated teon capillary excited with a 10 ns rise time pulse,232 most of these plasmas had been excited with relatively slow current pulses with rise times 50 ns. In these slow discharges, the large mass of wall-ablated material typically limited the electron temperature to values 60 eV. These plasmas were typically generated by discharging a low inductance capacitor 5100 nF through a circuit in which the capillary channel also acts as the main discharge switch with the purpose of minimizing the inductance. Capillary discharges of this kind have been employed to investigate gain by collisional recombination.66,151,152,233236 The capillary discharges that to date have proven to be most successful for soft x-ray laser excitation have a rapid current rise time, typically 1040 ns. The fast current rise time minimizes the amount of material that is ablated from the capillary walls before the magnetic eld compresses the plasma, detaching it from the walls.68,82,83 Diagnostics68,82,237,238 and modeling82,83,129,240,241 conrmed that the plasma of these fast capillary discharges is not a stationary homogeneous column, but that it rapidly contracts, heats up, and expands, constituting a kind of wall-inuenced Z pinch. These fast compressional capillary discharges have several demonstrated advantages to pump compact and efcient soft x-ray lasers: rst, the very high efciency with which they can create plasmas with high densities of multiply ionized species;68,69,239 second, the possibility to generate plasma columns with very large aspect ratios up to 1000:1 ;70,83 third, the very high axial uniformity that results from highly uniform initial conditions, a very rapid compression, and possible wallstabilizing effects,129 and fourth, the rapid motion of the plasma at the time of lasing, that causes a strong dynamic Doppler shift that helps to radiatively depopulate the laser lower level as a result of a reduction of the radial opacity.129 Figure 8 is an schematic representation of the pulsedpower generator and capillary discharge setup that was successfully used to amplify the 46.9 nm line of Ne-like Ar to intensities exceeding the saturation intensity.68,82,83 The capillary was placed in the axis of a 3 nF liquid dielectric ca-
FIG. 8. Schematic diagram of the capillary discharge setup used to obtain saturated amplication of the 46.9 nm line of Ne-like Ar from Rocca et al., Refs. 68 and 82 .
pacitor that was pulse charged by a Marx generator. The capillary loads were excited by discharging the capacitor through a spark-gap switch pressurized with SF6. The generator was used to excite Ar plasma columns up to 20 cm in length with current pulses of 39 kA peak amplitude and 75 ns rst half cycle duration.70 The same generator was also used to excite capillary channels of 1 mm diameter and 1 cm length with current pulses of 100 kA peak amplitude and 11 ns 10%90% current rise time.66 A more compact discharge setup based on a Blumlein transmission line that occupies a space of only 0.4 1 m2 in an optical table was used to produce laser output pulses with energy up to 25 J in the 46.9 nm line of Ne-like Ar.245 A photograph of this discharge is shown in Fig. 14 in Sec. V. More recently, a capillary discharge laser of similar size produced an average output pulse energy of 0.88 mJ at a repetition rate of 4 Hz at this wavelength.18 A higher power capillary discharge that generates current pulses of up to 200 kA with a 10%90% rise time of 10 ns was recently developed to explore the generation of the plasma conditions that are necessary for amplication at shorter wavelengths.242 The nal stage of the pulse generator consists of a radial water-dielectric Blumlein transmission line. The very rapid current rise time is obtained by switching the transmission line through an array of seven synchronized spark-gap switches. The high efciency with which these fast capillary discharges can generate highly ionized plasma columns is illustrated by the remarkable similarity of the two Ar spectra shown in Fig. 9.239 The spectrum in Fig. 9 a corresponds to a 43 kA, 13 ns rise time, 28 ns full width at half maximum FWHM discharge through a 2.5-mm-diam Ar-lled capillary channel,239 while the spectrum in Fig. 9 b is from a 1 MA current implosion in a large multiterawatt pulsed power machine Gamble II, see Ref. 243 . Argon plasmas driven by a 51 kA current pulse of similar characteristics in 1.5 mm diam capillaries were reported to reach an electron temperature of 150 eV.68 Moreover, spectra of similar discharges in CaH2 and TiH2 capillaries showed that the electron temperature is sufcient to excite the upper laser levels in Nelike Ca and Ti.69,239 The dynamics of the plasma column of a fast capillary discharge is illustrated in Fig. 10 by a sequence of timeresolved end-on soft x-ray pinhole camera images.237 The
X-ray lasers
3811
FIG. 9. Argon spectra corresponding to: a 1 MA current implosion in the Gamble II generator Ref. 243 and b 43 kA, 30 ns FWHM discharge current pulse through a 2.5 mm diam capillary Ref. 239 . Transition wavelengths are in .
FIG. 10. Sequence of time-resolved pinhole images of an Ar capillary discharge plasma viewed end-on. The capillary diameter is 4 mm and its length 12 cm, and the Ar pressure is 700 mTorr. The current pulse has an amplitude of 39 kA and a half period duration of 62 ns. The timing with respect to the origin of the current pulse is indicated. The images corresponding to times near maximum compression (9D),(E) were acquired with reduced detector sensitivity to avoid saturation from Tomasel et al., Ref. 237 .
images correspond to a plasma generated in a 4 mm diam polyacetal capillary lled with 700 mTorr of Ar by a current pulse of 62 ns rst half-cycle duration and 39 kA peak amplitude. During the rst part of the current rise time, the distribution of the current density and the soft x-ray incoherent plasma emission are localized near the capillary wall Fig. 10 a . Subsequently, the electromagnetic forces of the rising current pulse create a shock wave and rapidly compress the plasma to a size of about 300 m diameter Figs. 10 a 10 d . A signicant part of the current 20%50% is computed to ow through material ablated from the capillary walls by plasma radiation and heat conduction.83 The optimum conditions for lasing by collisional electron excitation occur several ns before stagnation, when the rst compression shock wave reaches the axis. Finally, the plasma column expands and cools Figs. 10 e 10 g . A second less signicant compression, that is of no interest for lasing, occurs later in time Fig. 10 h . Figure 11 illustrates the calculated evolution of the electron temperature and density in the vicinity of the axis of a similar Ar capillary discharge plasma excited by a current pulse having a half-cycle duration of about 70 ns. Lasing by collisional electron excitation of Ne-like Ar ions takes place at a time when the electron density is rapidly increasing and reaches (0.3 1 1019 cm 3), when the electron temperature is 6080 eV.83,241 Subsequently, the electron density continues to increase as the plasma stagnates, and laser action ceases due to increased refraction and collisional thermalization. The electron density and the spontaneous emission reach their peak values at stagnation 58 ns after the laser pulse.
V. SPECIFIC LASER SYSTEMS A. Capillary discharge lasers
1. Ne-like Ar laser
The rst observation of large soft x-ray amplication in a discharge-created plasma was realized in 1994 by Rocca et al. in the 3p 3s J 0 1 line of Ne-like Ar at 46.9 nm using the fast capillary discharge setup illustrated in Fig. 8.81,82 In that initial experiment, a discharge current pulse of 60 ns half-cycle duration and 40 kA peak current was used to excite Ar plasma columns in 4 mm diam capillary channels up to 12 cm in length. Figure 12 shows the measured increase of the intensity of the 46.9 nm line as a function of plasma column length, that corresponds to a gain-length product of gl 7.2 for the 12 cm long plasma columns. A small gain was also observed in the J 2 1 line of Ne-like Ar at 69.8 nm.82 The existence of small amplication in the J 2 1 line was later also reported by Hildebrand et al.244 Subsequent experiments employed longer plasma columns, up to 20 cm in length, under better optimized discharge conditions. Double-pass amplication experiments were also conducted using an Ir mirror.83 The latter experiments yielded an effective gain-length product of gl 27. These experiments resulted in the rst observation of gain saturation in a table-top soft x-ray amplier. The results of singlepass amplication measurements for capillary plasma columns up to 15.8 cm in length are shown as open circles in Fig. 13. The energy of the laser pulse is observed to increase exponentially for lengths up to about 12 cm, where it begins to saturate. A t of the data corresponding to plasma columns up to 11.5 cm with the Linford formula87 yields a gain
3812
J. J. Rocca
FIG. 11. Computed variation of the electron density and temperature on the axis of an Ar capillary discharge amplier. The capillary diameter is 4 mm, the Ar pressure 700 mTorr, and the current pulse has a half-cycle duration of 70 ns and a peak amplitude of 39 kA from Shlyaptsev et al., Ref. 241 .
coefcient of 1.16 cm 1. Saturation of the laser intensity was observed at gain-length products of about 14. The measured saturation behavior was found to be in good agreement with the result of two independent radiation transport models for the capillary plasma columns. The solid line in Fig. 13 is the result of calculations performed assuming parabolic gain and density proles, while the dashed line correspond to computations conducted for the time-dependent electron density and gain proles obtained from magnetohydrodynamic and atomic physics calculations.83 It should be noted that in these calculations consideration of refraction losses was found to be essential to adequately describe the measured energy dependence on plasma column length. The laser pulse width was measured to be about 0.8 ns.83 Based on the results summarized above, very compact saturated 46.9 nm lasers of size comparable to that of many widely utilized visible and ultraviolet gas lasers were developed.17,18,245 Figure 14 illustrates the size of one of these capillary discharge soft x-ray lasers in comparison to a 5 mW HeNe laser. The soft x-ray laser occupies a surface space of about 1 m by 0.4 m in an optical table.245 This capillary discharge pumped table-top laser was successfully used to perform high resolution soft x-ray laser interferometry and shadowgraphy of plasmas.246,274 Recently, a capillary discharge soft x-ray laser of similar size was operated at a repetition rate of 7 Hz to produce an average output pulse energy of 135 J, corresponding to an average laser power of 1 mW.17 The latest results include the generation of laser pulses with an average energy of 0.88 mJ at a repetition rate of 4 Hz in a highly saturated amplier.18 Figure 15 illustrates the laser output energy and average output power characteristics of this laser that used a ceramic capillary 34.5 cm in length. The spatially coherent average output power per unit bandwidth emitted by this compact table-top laser at 26.5 eV is comparable to that generated by a beam line at a third generation synchrothon facility.17 Its peak coherent power per unit bandwidth exceeds that of the synchrotron by nearly 6 orders of magnitude. With a peak spectral brightness of 1 1023 photons/ see mm2 mrad2 0.01% BW this table-top laser is among the brightest soft x-ray sources available.
FIG. 12. Axial spectra of Ar capillary discharge for three different plasma column lengths showing lasing in the 46.9 nm line of Ne-like Ar. A strong increase of the laser line intensity is observed from Rocca et al., Ref. 81 .
2. Ne-like Ar capillary discharge laser beam characteristics: Near-eld and far-eld patterns, and spatial coherence
The beam divergence and intensity distribution patterns from the Ne-like Ar capillary discharge laser were systematically measured over a wide range of discharge parameters.97 Figure 16 shows the variation of the two-
FIG. 13. Measured and computed 46.9 nm laser output energy as a function of capillary discharge plasma column length. Single and double-pass measurements are indicated by open and full circles, respectively. The lines are the result of two different radiation transport calculations from Rocca et al., Ref. 83 .
X-ray lasers
3813
FIG. 14. Photograph comparing the size of a saturated table-top 46.9 nm capillary discharge laser back to a 5 mW HeNe laser front from Colorado State University .
dimensional near-eld and far-eld intensity distribution of the laser output as a function of discharge pressure for discharges in a 4 mm diam polyacetal capillary. The size of the beam at the exit of the amplier and its divergence were observed to increase from about 150 to 300 m and from 2 to 5 mrad, respectively, as the pressure was decreased from 750 to 500 mTorr. Simultaneously, the beam intensity distribution changed from a single-peaked pattern to an annular prole. These effects were shown to be the result of increased refraction in the lower pressure discharges caused by larger electron density gradients.97 The larger gradients are mostly the result of a stronger compression and a reduced plasma column size in the lower pressure discharges. The spatial images of the beam intensity distribution corresponding to the higher pressure discharges are seen to have a very good cylindrical symmetry, and signicantly less structure than those that have been reported for laboratory-size laserdriven soft x-ray lasers.102 Another important laser beam characteristic is its degree of spatial coherence. Measurements in a capillary discharge pumped Ne-like Ar 46.9 nm laser observed for the rst time a monotonic increase of the spatial coherence as a function of plasma column length in a soft x-ray amplier,247 a result that is important for achieving a very good spatial coherence in cavity-less soft x-ray lasers. Several theoretical studies had predicted such improvement of the spatial coherence with amplier length as a result of gain guiding and refractive antiguiding that decrease the number of modes guided along an amplier column.248,249 In the capillary discharge pumped Ne-like Ar laser, the spatial coherence was measured recording the diffraction produced when the soft x-ray laser beam intersects a knife edge Fig. 17, top , a technique developed by Rus and rst utilized by Albert et al.250 to measure the coherence of a laser pumped Ne-like Zn laser. Figure 17 bottom shows the measured diffraction patterns corresponding to capillary plasma lengths between 8 and 16.4 cm, and compares them to the results of a wave optics model. The improvement of the coherence with amplier length is evident in the increased fringe visibility observed in the diffraction patterns corresponding to the longer plasma columns. Figure 18 shows results of a quantitative analysis of these data. The coherence is observed to increase mono-
FIG. 15. Measured output pulse energy and average output power of a table-top capillary discharge 46.9 nm laser operating at a repetition frequency of 4 Hz. a Shot-to-shot laser output pulse energy. b Average output power computed as a walking average of 60 contiguous laser pulses. c Distribution of the output pulse energy. The average pulse energy is 0.88 mJ and the standard deviation is 0.06 mJ from Macchietto et al., Ref. 18 .
tonically with capillary length, in good agreement with the result of the wave-optics calculations. For the longest capillaries studied, 16.4 cm, the coherence length in the tangential direction was determined to be about 4.5 mm in a plane situated at 5.9 m from the exit of the amplier. This corresponds to an effective coherence angle of 0.8 mrad, and to a beam that is approximately six times diffraction limited in the tangential direction. Off-axis measurements showed that
3814
J. J. Rocca
FIG. 16. Near-eld and far-eld patterns of a capillary discharge Ne-like Ar laser beam as a function of capillary discharge pressure. Diametral cuts with normalized intensities are shown at the left of each image. The measurements correspond to discharges in a 16.4 cm long, 4 mm diam polyacetal capillary excited by a 37 kA current pulse from Moreno et al., Ref. 97 .
the coherence length in the radial direction is 30%50% shorter than in the tangential direction.247 Wave-model computations suggest that a likely cause of the observed anisotropy of the spatial coherence is the change of the electron density during the laser pulse. The observed monotonic increase of the spatial coherence with plasma column length is an additional corroboration of the very high axial plasma uniformity that is obtained in fast capillary discharges.
3. Lasing in Ne-like S in vapor created by discharge ablation and scaling to shorter wavelengths
The capillary discharge Ne-like Ar amplication results were extended to Ne-like S by Tomasel et al., with the demonstration of a gain-length product of 7.5 in the 3s 1 P 0 3p 1 S 0 line at 60.84 nm.124 This experiment demon1 strated that large amplication in discharge created plasmas can be also obtained in elements that are solid at room temperature, a situation that will be frequently encountered in attempts to extend this scheme to shorter wavelengths. To obtain amplication in Ne-like S, the discharge setup illustrated in Fig. 8 was modied to allow for the injection of the sulfur vapor into the capillary channel through a hole in the ground electrode. The sulfur vapor was produced ablating the wall of an auxiliary capillary channel, drilled in a sulfur rod, with a slow current pulse delivering 200 J in about 50 s. The vapor generated by this capillary discharge was injected into the main capillary channel where it was subsequently excited by a fast current pulse having a rst halfcycle duration of 72 ns. Strong amplication was observed
FIG. 17. Top Schematic representation of the setup used to measure the spatial coherence of a Ne-like Ar capillary discharge laser. Bottom a Measured diffraction patterns corresponding to different plasma column lengths. b Cross sections of a . c Corresponding diffraction patterns computed using a wave-optics model from Marconi et al., Ref. 247 .
for a broad range of pressures, from 300 to 700 mTorr, and for currents between 33 and 38 kA. Figure 19 a shows a spectrum obtained under optimized conditions for amplication. The spectrum is completely dominated by the J 0 1 line transition of Ne-like S, which appears at 60.84
X-ray lasers
3815
FIG. 18. Variation of the spatial coherent length of a 46.9 nm capillary discharge Ne-like Ar laser as a function of plasma column length. The line is the result of a wave-optics model calculation from Marconi et al., Ref. 247 .
0.015 nm. Another line of Ne-like S, the 3p 1 P 1 3d 1 P 0 1 transition at 60.12 nm that in the absence of amplication should have similar intensity, also falls in the spectral range of Fig. 19 a . The fact that the intensity ratio of these two lines is observed to be at least 100 is clear evidence of amplication of the J 0-1 line. Figure 19 b shows the measured increase of the laser line intensity as a function of plasma column length. It corresponds to a gain coefcient of 0.45 cm 1, and to a gain-length product of 7.5 for 16.8 cm long capillaries. Lasing occurs shortly before stagnation of the plasma column, in a gain region of about 300 m in diameter as in the case of Ne-like Ar. The deceleration of the plasma column near the axis prior to stagnation results in a velocity gradient that, due to motional Doppler broadening, considerably facilitates the radial escape of the lower laser level radiation. At the time of lasing, the electron density and temperature in the gain region are computed to be about (2 3) 1018 cm 3 and 6080 eV, respectively. This temperature corresponds to a plasma that is overheated with respect to the temperature range of T e 20 40 eV for maximum Ne-like S abundance in a steady-state plasma. Due to the exponential dependence of the excitation rates on the electron temperature, the overheating causes a larger population inversion and, consequently, a larger gain than that attainable in steady state. Such overheating of the plasma with respect to steady-state ionization conditions can be more easily achieved in low-Z elements like S, due to a large decrease of the ionization time with ion charge. In addition, in this sulfur laser transient population effects are found to play a more important role than in the argon laser. While in these relatively long-lived discharge plasmas transient effects are not nearly as dramatic as in plasmas produced by picosecond lasers, the computations indicate that in the case of the discharge-pumped sulfur laser experiment transient population effects can increase the gain by 20%40%.124 Scaling of collisionally excited capillary discharge lasers to wavelengths shorter than 46.9 nm will require the use of elements heavier than Ar and higher excitation power densities. The scaling of plasma conditions and pump requirements for lasing at wavelengths as short as 10 nm have been
FIG. 19. a Spectrum of the axial emission of a capillary discharge in S. The amplied 60.8 nm 3s 1 P 0 3p 1 S 0 line of Ne-like S completely domi1 nates the spectrum, while the 60.12 nm 3p 1 P 1 3d 1 P 0 line of the same ion 1 that in the absence of amplication should have similar intensity is not observed. b Measured variation of the integrated line intensity of the 60.8 nm J 0 1 line of Ne-like S as a function of plasma column length from Tomasel et al., Ref. 124 .
estimated.82 In most cases, this will require the use of elements that are solid at room temperature. The necessary vapor density can be produced using discharge ablation or vaporization methods.69,124,239,251253 The same pulse generator used to saturate the J 0 1 line of Ne-like Ar was employed to conduct preliminary experiments in capillaries containing Ca and Ti. The spectra obtained yielded a clear identication of emission from the laser lines in Ne-Like Ca and a possible identication of the J 0 1 laser line in Nelike Ti.69 Recently, current pulses of up to 200 kA peak amplitude and 10 ns rise time were used to explore the generation of capillary plasma columns with the conditions necessary for amplication at shorter wavelengths.242
4. Gain in discharge-pumped recombination systems
The recombination scheme has the potential advantage of allowing for a more rapid scaling of discharge-pumped ampliers to shorter wavelengths. Evidence of gain has been reported in several discharge-pumped recombination laser experiments.151155,66 However, in all cases only small exponentiation (gl 4) has been observed to date, with the laser line intensity remaining of the same order as that of sur-
3816
J. J. Rocca
FIG. 20. Spectra of the axial emission of an oxygen Z-pinch plasma as a function of plasma column length. A nonlinear increase of the intensity in the 52 nm 4 f 3d line of Li-like O line is observed during plasma recombination from Wagner et al., Ref. 153 .
rounding nonlasing lines emitted by the plasma. Such a small increase in the line intensity complicates the interpretation of the results and the verication of gain in recombination laser experiments. This is because in recombination systems the laser lines can be intense even when no amplication is present, and phenomena such as end effects that affect the axial homogeneity of the plasma parameters can inuence the intensity dependence with plasma column length. The possibility of obtaining amplication by collisional recombination in capillary discharge plasmas was proposed by Rocca et al. in 1988, who analyzed the case of amplication at 18.2 nm in the 3-2 line of H-like C.228 Early experiments by Marconi and Rocca conducted in a LiH capillary observed an increase in the intensity of the 3-2 line of H-like Li in the recombination phase of a plasma.233 Similar results in H-like Li were later obtained by Pockl et al., who also conducted gain experiments in Li-like 0.254 Several subsequent experiments were conducted in capillary discharges with the goal of obtaining amplication in the 18.2 nm line of C VI.66,151,152,234 Steden and Kunze reported as evidence of gain the observation of short spikes in the intensity of the 18.2 nm line of H-like C, occurring at the time of maximum current during the second half-cycle of the current.151,255 The experiments were conducted using polyacetal capillaries excited by a low-energy discharge pulse of 240 ns period. The intensity of the spikes increased with the capillary length, although not exponentially as it would be expected for the case of a constant gain coefcient.151 Attempts to observe
these spikes in similar setups at other institutions were unsuccessful,234236 and spectroscopic studies suggested that the plasma conditions do not satisfy those required for the observation of amplication by collisional recombination in H-like C.235,236 Kunze et al. later suggested that the spikes could be amplied spontaneous emission due to charge exchange between C VII ion streams created in the neck region of an m 0 instability and C III and C V ions present in the colder surrounding regions.255 Experiments conducted at Colorado State University showed that a much stronger 18.2 nm C VI emission can be obtained in polyethylene capillaries.66,234 Polyethylene capillaries excited by fast current pulses 11 n rise time, 40 ns half-cycle duration yielded time-resolved spectra in which anomalously large intensity ratios between the 18.2 nm 3-2 and 13.5 nm 4-2 lines of H-like C are indicative of amplication.66 However, these experiments did not show a reproducible supralinear increase of the laser line intensity with capillary length. Shin et al.152 subsequently reported evidence of amplication of the 18.2 nm H-like C line in polyethylene capillaries 1.2 mm in diameter excited by a current pulse of 220 ns half-cycle duration, generated discharging a 120 nF capacitor charged to 1727 kV. Time-integrated spectra showed an exponential growth of the laser line intensity for capillary lengths between 0.8 and 1.6 cm, corresponding to a gain coefcient of 2.8 cm 1. The supralinear increase of the line intensity was observed to level off for plasma lengths longer than 1.6 cm, and to be present only in the rst few discharges of newly prepared capillaries. Recently, Dussart et al. conducted experiments in a similar ablated polyethylene capillary discharge and reported the observation of a small superlinear increase in the C VI Balmer- and - lines.152 Gain following plasma recombination has also been reported in several Z-pinch plasmas. Experiments performed by Glenzer and Kunze utilizing a gas-liner plasma pinch yielded evidence of gain in the 4 f 3d lines of O VI and F VII at 52.0 and 38.2 nm, respectively.154 Measurements of amplication as a function of plasma length are difcult to perform in this device, and were not reported. Instead, the gain was inferred from the ratio of axial and radial line intensities. For O VI this ratio was observed to be indicative of a gainlength product of 4.5. In the case of F VII measurements were only conducted in the radial direction, and the existence of line enhancement was inferred from the ratio of the intensities of the 4 f 3d and 4d 3p lines. The strongest amplication was reported to occur always at times shortly before maximum pinch compression. Pinhole pictures of the plasma revealed the existence of RayleighTaylor instabilities. Collisional recombination was suggested as the mechanism most probably responsible for the observed line enhancements, but a possible inuence of instabilities was not ruled out.154 Wagner et al. obtained axial emission spectra of an oxygen z pinch that show a small supralinear increase of the intensity of the 4 f 3d and 4d 3p lines of Li-like O at 52.0 and 49.8 nm as a function of plasma length up to 9 cm.153 From these data, shown in Fig. 20, gain-length products of about 2.5 and 2.2 were reported for each of these two lines, respectively. The supralinear increase of the intensity of these lines occurred about 45 ns after the maximum compression, during
X-ray lasers
3817
the expansion of the plasma. This time delay is consistent with the interpretation of amplication by collisional recombination. Amplication experiments in the isoelectronic equivalent of these lines in Li-like Ne at 29.2 nm and 28.2 nm were performed by Boss et al.155 Discharges in 6 mm diam tubes yielded two uorescence maxima in these lines, the second corresponding to excitation by recombination of He-like Ne ions. The observation of a relative increase of the second intensity maxima with respect to the rst in longer plasma columns, and the narrowing of its temporal prole were reported as evidence of gain. However, a measurement of exponential growth of the intensity of these laser lines with plasma column length has not yet been reported. In summary, several recombination laser experiments in discharge-created plasmas have yielded indications of amplication. However, in all the experiments the maximum amplication obtained to date has been small. Additional efforts are required to increase the amplication to the level necessary for the development of a practical laser.
B. Laser pumped by ultrashort pulse lasers
FIG. 21. Schematic illustration of the gas target setup used to obtain lasing in Pd-like Xe at 41.8 nm in a plasma created by optical eld induced ionization. Excitation was provided by circulary polarized laser pulses with an energy of 70 mJ and a duration of 40 fs from a CPA Ti:sapphire laser operating at 10 Hz from Lemoff et al., Ref. 75 .
1. Collisionally excited Pd-like Xe laser in an opticaleld-ionized plasma
Lemoff et al. realized in 1994 the rst demonstration of a collisionally excited laser in an OFI plasma, obtaining a gain of gl 11 12 in the 41.8 nm line of Pd-like Xe.75 In this experiment, an intense circularly polarized femtosecond optical laser pulse was used to simultaneously create the highly ionized lasing ions and the hot pumping electrons. In this excitation method, originally suggested by Burnett and Corkum,256 tunneling ionization strips the atoms to create the ground-state ions in the ionization stage of interest, and produces electrons with sufcient energy to collisionally excite these ions to the upper laser level. Lemoff, Barty, and Harris proposed the implementation of this scheme in eight time ionized gases to obtain amplication in lines in the 3050 nm wavelength region.257 In the successful Pd-like Xe experiment the excitation was provided by circularly polarized laser pulses with an energy of 70 mJ and a duration of 40 fs generated by a CPA Ti: sapphire laser operating at 10 Hz. The radiation of the pump laser was longitudinally focused onto a Xe gas target to tunnel ionize the atoms to the Pd-like ionization stage and to simultaneously produce the hot electrons necessary to collisionally excite the laser upper level. Figure 21 illustrates the differentially pumped gas target used in the experiment. The gas target was implemented by continuously admitting Xe gas into a cell provided with 500 m diam pinholes for the entrance and exit of the pump laser beam and for the exit of the amplied soft x-ray radiation. The gas cell was placed inside a main vacuum chamber that was continuously pumped to maintain a pressure of less than 2 10 3 Torr. In this way, the femtosecond laser pulses were allowed to propagate in a nearly gas-free region until immediately before entering the gas cell that contained up to 12 Torr of Xe. The pinholes were placed on translation stages to conveniently change the gas target length, allowing for the measurement of the amplication as
a function of plasma length. The pump laser was circularly polarized by a mica quarter waveplate and was focused by a 50 cm focal length mirror to an intensity 3 1016 W/cm2 over a length of at least 7.4 mm. Figures 22 a and 22 b show spectra corresponding to Xe target pressures of 3 and 12 Torr respectively. The 12 Torr spectrum shows an increase of more than 2 orders of magnitude in the recorded integrated intensity of the 41.8 nm laser line with respect to the 3 Torr spectra, a clear indication of amplication. A more direct measurement of the gain was obtained from the variation of the intensity of the laser line as a function of gas target length. The result of this measurement is shown in Fig. 23. The gain coefcient was estimated at 13.3 cm 1 using the Linford formula,87 corresponding to a gain-length product gl 11.2 for an estimated target length of 8.4 mm. When the range was limited to the data from 3.9 to 7.4 mm dashed line , a better t was obtained with a gain coefcient of 16.8 cm 1 and gl 12.4 for a 7.4 mm target. Subsequent experiments determined that to achieve a high gain it is critical to avoid a signicant prepulse in the femtosecond laser pulse.258 The laser pulse width was not measured. However, at the operating pressure of 12 Torr, the 16.5 nm line radiation that links the laser lower level to the ground state is trapped, with an estimated absorption length of less than 0.1 m. This forces the stimulated emission to self-terminate after a time interval on the order of the upperstate lifetime, estimated at about 3.4 ps at 12 Torr.75 This collisionally excited OFI laser has the advantage of requiring a low excitation energy, allowing for high repetition rate operation. Amplication at a repetition rate of 10 Hz was already demonstrated in the successful proof-ofprinciple experiment conducted at Stanford.75 A remaining challenge in the development of this laser scheme is the further increase of the amplication required to generate the practical output pulse energies that will enable its use in applications. This will most likely require a method to channel the tightly focused pump laser beam over longer distances using a waveguide. Scaling of this scheme to shorter wavelengths will require higher intensities of the pump pulse. For example, lasing in Ni-like Xe at 10 nm will require a laser intensity of about 5 1018 W cm 2. Alternatively, it has been proposed that this scheme could be scaled
3818
J. J. Rocca
FIG. 24. a Schematic illustration of the experimental setup used at RIKEN in the observation of amplication in the 13.5 nm line of H-like Li in a plasma created by optical eld induced ionization from Nagata et al., Ref. 72 .
FIG. 22. Data showing the amplication of the 41.8 nm line of Pd-like Xe in a plasma created by optical eld induced ionization; a and b are spectra corresponding to Xe target pressures of 3 and 12 Torr, respectively. c Variation of the integrated intensity of the 41.8 nm line of Pd-like Xe as a function of gas target length from Lemoff et al., Ref. 75 .
to shorter wavelengths utilizing Be-like ions.259

2. Ampliers based on recombination of an opticaleld-ionized plasma
Amplication has been observed at 13.5 nm in the 21 transition of H-like Li in recombining plasmas ionized by a strong optical eld.72,74,260,261 The gain obtained in different experiments ranges from gl 4 Ref. 72 to gl 6.5.260 The rst observation of amplication in an OFI plasma was reported in this transition by Nagata et al.72 The setup of the initial OFI recombination laser experiment conducted at the RIKEN Japan in 1993 is schematically illustrated in Fig. 24. A cold plasma of totally ionized Li ions was generated in two steps. First, a plasma of singly ionized Li ions was produced by line focusing a 20 ns KrF laser pulse with an energy of 200 mJ onto a rotating lithium target located in an evacuated chamber. Second, after a selected time delay, the plasma was optical-eld ionized with a pulse of 50 mJ en-
FIG. 23. Variation of the integrated intensity of the 41.8 nm line of Ni-like Pd as a function of Xe gas target length in a plasma created by optical eld induced ionization from Lemoff et al., Ref. 75 .
ergy from a linearly polarized subpicosecond 0.5 ps KrF laser pulse. The subpicosecond laser pulse was longitudinally focused to an intensity of 1 1017 cm 2, at a distance of 0.5 mm from the target surface. The Lyman series emission from the H-like Li ions was detected with a at eld grazing incidence spectrograph. Figure 25 shows the measured variation of the 13.5 nm laser line intensity with plasma length, from which a gain coefcient of 20 cm 1 was deduced. The 13.5 nm radiation pulse was measured to have a FWHM duration of less than 20 ps, with the peak intensity occurring about 20 ns after the pump laser pulse.261 This delay suggests that the laser upper level is not directly populated by the pump laser, but by collisional relaxation processes following collisional recombination. However, since the maximum plasma length was only 2 mm, the resulting gain-length product was limited to gl 4. Similar experiments conducted at Princeton also using a KrF laser pump,179 and at Berkeley using the second harmonic of a Ti:sapphire laser178 conrmed the nonlinear growth of the intensity of the Lyman transition of H-like Li observed at RIKEN. Nevertheless, in all three experiments the maximum amplication length was limited to about 2 mm by the difculty of propagating the radiation in the plasma, due to ionizationinduced refraction. This is a limitation of the confocal geometry that, as a result of maximum intensity on axis, induces the highest degree of ionization and electron density on-axis. Approaches to overcome this problem have been suggested, and include the generation of very at transverse intensity proles,178 and the generation of plasma waveguides with a minimum density on axis.74,119121 Progress in the implementation of the latter approach was reported by Korobkin et al.,74 who used the setup illustrated in Fig. 26 to increase the amplication length to 5 mm by preforming a plasma waveguide in a LiF microcapillary tube. A plasma with a minimum electron density on-axis was generated by ablation of the capillary walls with a 100 mJ, 5 ns pulse of 1.06 m wavelength radiation from a Nd/YAG laser loosely focused on the entrance of the microcapillary. Following a time delay of several hundred nanoseconds, the plasma was further ionized with a laser pulse of 50 mJ energy and 250 fs duration from a KrF excimer laser that was tightly focused at the entrance of the microcapillary. The observed increase of the intensity of the 13.5 nm line of H-like Li as a function of length is shown in Fig. 27, and corresponds to gl 5.5.74 These results were obtained at a repetition rate of 2 Hz.
X-ray lasers
3819
FIG. 25. Results of amplication experiment in the 2-1 line of H-like Li in a plasma created by optical eld induced ionization. Intensities of the 13.5 nm line of H-like Li and 19.9 nm line of He-like Li as a function of plasma column length. The dashed curve is a calculated increase corresponding to a small-signal gain coefcient of 20 cm 1 from Nagata et al., Ref. 72 .
Gain by recombination in an OFI plasma has also been reported in other ions. Chichkov et al. investigated amplication in transitions to the ground states in low-charge ions: O II, O III, C III, and N III.262 A motivation for the study of amplication in these ions was their larger ratio of statistical weights between the lower and upper laser levels (g 1 /g 2 1), which is more favorable for the generation of population inversion. The ions of interest were produced by OFI of O2, CO2, and N2 gases in gas jets using 80 mJ pulses from a 150 fs Ti:sapphire laser. Evidence of amplication was observed in the 2p2s 3 P 2p 2 3 P transition of O III ( 37.4 nm) and in the 2p 2 3s 2 P 2p 3 2 D line of O II ( 61.7 nm). The data suggest gain coefcients as high as 25 cm 1, but again the overall amplication was small due to the small length of the gas targets. Recently, Chichkov et al. have questioned the interpretation of the H-like Li OFI amplication experiments based on the possible existence of Li clusters.175 However, additional experiments are needed to conrm this hypothesis. In summary, the OFI laser recombination scheme in transitions to the ground state has been reported to generate amplication at wavelengths as short as 13.5 nm with laser excitation energies of only 50 mJ. The small excitation energies are compatible with the requirement for the generation of high repetition rate table-top soft x-ray lasers. However, the remaining challenge in realizing a practical OFI laser is the increase of the gain-length product from the reported values of 6.5 to the larger values necessary to achieve practical output energies. Predictions of the output pulse energies that can be expected of OFI lasers range from 1 J Ref. 178 to 1020 J.260 The OFI recombination scheme can in principle also be scaled to shorter wavelengths by using ions with higher ionization potentials. This could be implemented using more intense optical elds, provided that parametric electron heating effects mainly stimulated Raman backscattering are not found to be a limitation.
3. Transient inversion collisional lasers The recent availability of relatively compact multiterawatt ultrashort-pulse optical laser systems with output ener-
FIG. 26. Schematic illustration of the experimental setup used at Princeton in the creation of an OFI plasma in LiF microcapillaries for the observation of amplication in the 13.5 nm line of H-like Li from Korobkin et al., Ref. 74 .
gies of several Joules has opened the opportunity to demonstrate soft x-ray lasing by transient electron collisional excitation.7678,263 This approach has allowed researchers to reduce the size of high gain laser-pumped soft x-ray collisional lasers from that of fusion-type laser facilities to the space occupied by a few optical tables. The implementation of transient inversion lasers is based on a two-step excitation
FIG. 27. a Axial spectra of an OFI plasma created in a LiF microcapillary using the setup shown in Fig. 26. The n 2 1 transition of H-like Li at 13.5 nm is observed to increase as a function of plasma length with a gain coefcient of 11 cm 1 from Korobkin et al., Ref. 74 .
3820
J. J. Rocca
sequence. First, a long scale length plasma containing the active ions must be prepared, since the inversion lifetime is usually shorter than the ionization time. This also allows for the elimination of the micron-scale plasma density gradients that are created when a solid target is irradiated with a single ps pulse. The experiments reported to date have used a rst laser pulse with a duration on the order of 1 ns and energy of several Joules.7678 Second, after a time delay selected to optimize the degree of ionization, electron density, and density gradients, the plasma is rapidly heated with a picosecond or subpicosecond laser pulse. The ultrashort pulse rapidly increases the electron temperature to match or exceed that corresponding to the excitation energy of the laser upper level, generating the transient inversion. The rst demonstration of amplication by transient inversion was realized by Nickles et al. in the 32.6 nm 3p 3s J 0 1 line of Ne-like Ti.76 The experiment was performed at the Max Born Institute using the hybrid CPA Ti-sapphire/ Nd:glass pump laser diagrammatically illustrated in Fig. 7. This laser system delivers synchronized long and short laser pulses of 1.2 ns and 0.7 ps duration, respectively. A dual cylindrical lens optics system was used to line focus the two beams onto the target with a width of 30 m over lengths of 15 mm. The diagnostics system consisted of a transmission grating and a streak camera. No laser pulse was observed as long as only one of both laser pulses was used. The rst observation of gain was realized with short pulses of more than 2 J and long pulses of more than 3 J. Figure 28 a shows a streak camera record of the spectrally and temporally resolved plasma emission in the 1050 nm spectral region for a 5 mm long plasma created with pump intensities of 1012 and 1015 W/cm2 for the long and short pulse, respectively. The bright and short spot at the top corresponds to strong amplication of the 32.6 nm of Ne-like Ti. The soft x-ray laser pulse was measured to have a duration of less than 20 ps. Weaker lasing was also observed in a line near 30 nm. This second laser line was identied as the 3d 3p J 1 1 transition of Ne-like Ti. The inversion in this line is made possible by the reabsorption of the 3d 1 P 1 2p 1 S 0 transition.192,193 Figure 28 b illustrates the measured growth of the intensity of the 32.6 nm laser line as a function of plasma column length, for plasmas as long as 5 mm. Despite the large statistical uctuations a nonlinear increase is clearly observed. It was analyzed to correspond to an average gain of 19 cm 1 and to a gain-length product of 9.5.76 This gain coefcient is about seven times larger than that previously reported for the same line in the quasisteady state regime when pumped by a 0.6 ns duration pulse with an energy of 200500 J.103 These transient inversion Ne-like Ti results were reproduced and further improved by Dunn et al. at Lawrence Livermore National Laboratory LLNL , who reported a gain-length product of 14 in this laser line and also obtained amplication at 25.5 nm using Ne-like Fe.78 The size of the system utilized by Dunn et al. is that illustrated in Fig. 29, which depicts the layout of the Compact Multipulse Terawatt laser facility COMET at LLNL. The pump laser occupies two 1.2 m 3.6 m optical tables with a total area of less than 10 m2. It is a 15 TW hybrid laser system consisting of a Ti:sapphire oscillator tuned to 1050 nm wavelength and
FIG. 28. a Results of transient inversion soft x-ray amplication experiment in Ne-like Ti. Spectrally resolved streak from a 5 mm long Ti plasma produced as the result of a sequence of a long 1.5 ns, 5 J and short 0.7 ps, 2.6 J laser pulses. The intense, short duration emission at 32.6 nm corresponds to the 3p 3s J 0 1 line of Ne-like Ti and is the result of strong amplication. b Nonlinear growth of the 32.6 nm laser line intensity with plasma lengths of 2, 3, and 5 mm from Nickles et al., Ref. 76 .
four-stage Ndphosphate glass ampliers. It generates two synchronized laser pulses of 7.5 J in 500 fs and 15 J in 600 ps at a repetition rate of one shot every 4 min. In some of the experiments the compressor gratings were detuned to produce 1 ps pulses. Typically 50 shots on target are achieved in a day. Transient collisional excitation experiments conducted at Rutherford Laboratories with 15 J of short-pulse excitation energy from the Vulcan laser facility saturated the 32.6 nm line of Ne-like Ti and the 19.6 nm of Ne-like Ge.263 While this pump laser is not a table-top system, a measurement of the laser output intensity versus pump energy suggested that saturation of the Ne-like Ti line could also be achieved with a short-pulse excitation energy of 5 J. The soft x-ray laser output energy was measured to be 20 J, corresponding to an efciency of 2 10 6 . The transient collisional scheme was also extended to shorter wavelengths utilizing the Ni-like sequence.77,78 Dunn et al. reported a gain of up to 35 cm 1 and a gain-length product of 12.5 in the 4d 1 S 0 4p 1 P 1 line of Ni-like Pd at 14.7 nm. In this experiment, the plasma was formed irradiating a slab target with a sequence of two laser pulses: a 5 J pulse of 800 ps duration followed by a 5 J pulse of 1.1 ps duration. The spectrum in Fig. 30 a shows the increase of
X-ray lasers
3821
FIG. 29. Schematic illustration of the CPM pump laser system and target chamber used to produce transient inversion collisional soft x-ray lasers at Lawrence Livermore National Laboratory. The pump laser occupies two standard optical tables of dimensions 1.2 m 3.6 m with total area less than 10 m2. The system can be red once every 4 min Courtesy of J. Dunn, LLNL .
the 14.7 nm line intensity observed when the target length was increased from 1 to 3 mm. The measured increase of the integrated laser line intensity for target lengths up to 8 mm is shown in Fig. 30 b . The gain coefcient for lengths between 1 and 2 mm is about 35 cm 1. The gain coefcient continuously decreases to reach a value of 3.9 cm 1 for target lengths above 7 mm. The overall gain-length product was estimated at 12.5 by continuously calculating the gain and integrating it over the full 8 mm target length. This smooth decrease of the gain with target length has been observed in all transient collisional excitation experiments and resembles gain saturation. However, it is mainly a consequence of transient time effects that cause a gradual decrease of the gain as the x-ray laser beam propagates through the plasma. Model computations suggest the transient gain is very inhomogeneous in space and time, reaching values of 200 cm 1 during the rst 2 ps and rapidly decreasing to 28 cm 1 by 10 ps Fig. 30 c .77 The highest gain occurs in a thin layer near the critical density region while the lower gain is computed to occur in areas with a relatively at, lower density prole at 100300 m from the target surface. The short inversion lifetime decreases the photon effective amplication path in the high gain region to an axial length of only 500 nm. The model computations also suggest that refraction effects caused by density gradients normal to the target surface shift the x-ray laser beam out of the high density and high gain region for longer target lengths, contributing to the gradual decrease of the gain.77 As a consequence of these effects, the lower density regions are more adequate for amplication, in spite of the smaller local gain. Most recently the transient inversion scheme was extended to several Ni-like ions ranging from Y ( 24.01 nm) to Mo ( 18.89 nm) producing strong lasing on the 4d 1 S 0 4p 1 P 1 lines.216 It can be expected that a large number of collisionally excited laser tran-
sitions will be demonstrated to lase in similar systems in the near future. In summary, the transient excitation of preformed plasmas with high power ultrashort-pulse lasers has reduced by nearly 2 orders of magnitude the pump energy required to excite high-gain laser-pumped collisional soft x-ray lasers in Ne-like and Ni-like ions. This has greatly reduced the size of these lasers and has increased their repetition rate up to one shot every few minutes. It has been suggested that transient inversions driven with ultrashort-laser pulses with energies on the order of 1020 J will lead to the realization of high gain x-ray lasers well below 10 nm, possibly reaching wavelengths within the water window 2.34.4 nm region of the spectrum.76 Future improvements in pumping efciency resulting from optimized target congurations and traveling wave excitation can be expected to lead to saturated transient inversion soft x-ray lasers occupying a single optical table. See notes added in proof, Dunn et al.
C. Lasers pump with long pulse lasers 50 ps10 ns
While in the last few years the majority of the efforts toward the development of laser-pumped table-top soft x-ray lasers has shifted to the use of ultrashort-pulse pump lasers,72,7478,216 signicant gain has also been reported from experiments that used longer pump pulses, in the range of 50 ps5 ns.73,79,80,264272 Recombination,73,80,264269 and collisional excitation schemes79,144,270272 excited by single or multiple laser pulses with pulse width ranging from 50 ps to 10 ns have been studied. A series of experiments were conducted during the last 10 years at Princeton University toward the development of a small-scale recombination laser in the 18.2 nm line of C VI utilizing compact Nd-YAG lasers as the pump source.264267
3822
J. J. Rocca
FIG. 30. Results of transient inversion soft x-ray amplication experiment in the 4d 4p J 0 1 line of Ni-like Pd at 14.7 nm. a Spectra showing the increase of the intensity of the laser line as a function of target length. The intensity increases by more than 3 orders of magnitude when the target length is increased from 1 to 3 mm. b Measured intensity of the laser line as a function of target length. The continuous line is a guide through the experimental points. The data show the decrease of the transient gain as the x-ray laser propagates along the plasma column. c Computed transient gain proles as a function of time measured from the arrival of a 1 ps excitation pulse from Dunn et al., Ref. 77 .
In 1989 Kim et al. reported a gain of about 8 cm 1 in that line as a result of an experiment that used a 25 J pump pulse of 3 ns duration, line focused into a 5 mm long solid carbon target.264 In this experiment, a stainless steel blade was placed in proximity with the target to provide additional cooling by thermal conduction and by radiation from iron impurities. Later experiments obtained similar gain coefcients in 6 mm long targets with only 4 J of laser excitation energy using a multin carbon target.266 However, difculties were found in increasing the amplication beyond a gain-length products 5. The latest attempt in increasing the gain-length product in this line utilized a polyethylene microcapillary as target.267 The Nd-YAG laser 2.5 J, 1.5 ns
pulse width in this experiment was focused at the entrance of the microcapillary with the motivation of achieving guiding of the pump beam, plasma connement, and plasma cooling by heat conduction to the walls. A gain-length product of about 5 was inferred from the intensity ratio of lines in the axial and perpendicular direction the latter through small holes in the microcapillary walls . However, shot-to-shot intensity uctuations prevented the verication of this result by the measurement of the increase in the laser line intensity with plasma column length.267 In contrast, a similar experiment recently conducted in a B2O3 microcapillary succeeded in observing amplication up to gl 5 in the 32 transition of H-like B at 26.2 nm.80 The most remarkable aspect of this experiment is that pumping was done at 1 Hz with two low power lasers occupying a space of 1.2 3 m in an optical table. The setup resembled that of the OFI H-like Li recombination laser experiment illustrated in Fig. 26, with the main difference that a 0.4 J, 8 ns Nd/YAG laser replaced the ultrashort-pulse KrF laser, making the arrangement considerably more compact. The preplasma was created by a 20 ns, 0.2 J KrF laser. The measured growth of the H-like B 26.2 nm line as a function of plasma length is shown in Fig. 31. Gain in this transition had been previously reported by Goodberlet et al.73 who conducted table-top recombination laser experiments utilizing a shorter excitation pulse of 60 ps duration with an energy of about 0.7 J. A gain-length product of about 4 was reported in the 26.2 nm line of H-like B utilizing a target conguration with a cooling blade, similar to that used in some of the C VI experiments at Princeton. However, the authors later realized that the streak camera used in the experiments had a nonlinear intensity response, a fact that creates an uncertainty in the results.273 Hara et al. realized the rst experiments that utilized multiple pulse excitation from a pulse train to excite a recombination laser.268 They have conducted a lengthy series of studies with the objective of developing a compact soft x-ray recombination laser based on a train of 100 ps pulses produced by a tabletop Nd-YAG laser system.71,268,269 In this case, the motivation for the use of a pulse train is a more efcient production of highly ionized ions and a faster termination of the excitation, determined by the dropoff of the last pulse. Experiments conducted utilizing 1.52 J of laser pulse energy to excite an 11 mm long Al slab target were reported to produce a gain coefcient of 3.2 cm 1 in the 15.5 nm line of Li-like Al.71 The latest results include double-pass experiments and multipass experiments that made use of unstable laser resonators constructed using Mo/Si multilayer mirrors. Evidence of cavity enhancement in the 15.5 and 15.1 nm lines of Lilike Al was observed.269 Signicant efforts have also been devoted to reduce the size of collisional lasers using laser-pumped pulses of 50 500 ps duration. In 1988 Hagelstein138 proposed to scale down the size of collisionally excited lasers by exciting low-Z Ni-like ions with 12 J pump pulses and predicted substantial transient gains. An experiment was implemented to demonstrate gain near 20 nm using a series of 60 ps laser pump pulses.270,271 This initial attempt to scale down the size of laser-pumped collisionally excited soft x-ray lasers to a single optical table was also one of the rst experiments to
X-ray lasers
3823
tive study of the inuence of the Pd target surface roughness on the laser output performance indicated a substantial increase in laser output intensity for reduced roughness. Presently, the repetition rate of these soft x-ray lasers is about one shot every 30 min limited by the frequency of ring of the pump laser.
VI. DISCUSSION AND FUTURE PROSPECTS
FIG. 31. Result of amplication experiment in the 3-2 line of H-like B ( 26.2 nm) using a plasma created by a sequence of two low power laser pulses in a B2O3 microcapillary plasma. The on-axis spectra for plasma lengths of 1, 2, and 3.5 mm show the nonlinear increase of the 26.2 nm laser line from Korobkin et al., Ref. 80 .
use multiple-pulse excitation. The pump pulses were generated by a laser consisting of a mode-locked Q-switched Nd:YLF laser, followed by a preamplier and a slab amplier. Line-focus excitation of a Nb target was reported to result in gains of about 3 cm 1 and gl 3 in the 20.4 nm line of Ni-like Nb. Similar experiments using the neighboring elements identied emission from the candidate laser line in Ni-like Mo 18.9 nm and Ni-like Zr 22.1 nm .271 More recently, Balmer et al. have reported the observation of large amplication in Ne-like ions irradiating slab targets with 100 J pulses of 500 ps duration produced by a Nd:glass laser that occupies a laboratory area of about 9 4 m. 79,272 These experiments made use of the prepulse technique that had been successfully used to increase the efciency of collisional lasers excited by laboratory-size laser pumps.103117 The 3p 3s J 0 1 lines of Ti, Cr, Fe, Ni, and Zn were amplied to reach an intensity at least an order of magnitude larger than the background radiation using a prepulse with 0.7% of the total laser pump energy that preceded the main pulse by 5 ns. For the 32.6 nm line of Ti, the 25.5 nm line of Fe, and the 23.1 nm line of Ni gain coefcients between 4.2 and 3.6 cm 1 were obtained for 2.4 cm long curved targets, resulting in gain length products of about 10.79 Angleresolved spectra indicated a typical beam divergence of about 3 mrad FWHM. At a recent conference, the same group has reported saturation of the 25.5 nm line of Ne-like Fe and also saturation of the 14.0 and 14.7 nm lines of Nilike Ag and Ni-like Pd, respectively.144 These latest results in Ni-like ions were obtained using a sequence of a single 8% prepulse followed after 1 ns by a main excitation pulse of 30 J, 100 ps. Gain-length products of 15.5 and 15.6 were obtained on at Ag and Pd targets, respectively. A compara-
Remarkable progress has been achieved in the last few years in several critical aspects of the development of practical table-top soft x-ray lasers. It includes advances in both table-top laser-pumped and discharge-pumped devices. Laser-pumped soft x-ray lasers have reached saturation at wavelengths as short as 14 nm using only several Joule of excitation energy. Signicant soft x-ray amplication has also been observed in several systems utilizing only a fraction of a Joule of laser excitation energy. A very compact 26.5 eV discharge-pumped laser has been operated at a repetition of several Hz producing millijoule-level laser pulses and an average power of 3.5 mW. This table-top laser produces a spatially coherent laser average power per unit bandwidth that is comparable to that generated by a thirdgeneration synchrotron beam line, and a peak coherent power that is several orders of magnitude larger. The reduction in the size, cost, and complexity of soft x-ray lasers resulting from these developments will soon lead to the widespread use of these sources in numerous scientic and technological applications. Some table-top soft x-ray lasers have already emerged from their development stage to be used in applications. For example, a capillary dischargepumped laser has been successfully utilized to perform soft x-ray plasma shadowgraphy and interferometry.246,274 These experiments will soon be followed by an explosion in the number of applications of table-top soft x-ray lasers. Signicant progress can also be expected in the next few years in the demonstration of compact practical lasers at increasingly shorter wavelengths as a result of the combination of compact high power ultrashort-pulse laser systems with transient inversion collisional schemes, or recombination lasing in transitions to the ground state. Moreover, the advent of relatively compact optical laser systems with output powers of up to 100 TW can be expected to result in the long awaited realization of K-shell photoionization lasers operating at extremely short wavelengths. Notes added in proof. Since the submission of this Review Article, the proceedings of the latest International Conference on X-Ray Lasers have been published. X-Ray Lasers 1998 Proceedings of the 6th International Conference on X-Ray Lasers, edited by Y. Kato, H. Takuma, and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 . The proceedings of another recent conference Soft X-Ray Lasers and Applications III, edited by J. J. Rocca and L. B. Da Silva, SPIE Society of Photo-optical Engineers, Bellingham, WA, 1999 , Vol. 3776, are in press. At the recent conference Soft X-Ray Lasers and Applications III, Denver, CO, 1823 July 1999, Dunn et al. reported saturation of the 14.7 nm Ni-like Pd and 14.0 nm Ni-like Ag transient inversion collisional excitation lasers
3824

25
J. J. Rocca A. V. Vinogradov, I. I. Sobelman, and E. A. Yukov, Sov. J. Quantum Electron. 7, 32 1977 . 26 L. A. Vainshtein, A. V. Vinogradov, V. I. Safranova, and I. Vu. Skolev, Sov. J. Quantum Electron. 8, 239 1978 . 27 A. V. Vinogradov and V. N. Shlyaptsev, Sov. J. Quantum Electron. 10, 754 1980 , and 13, 303 1980 . 28 W. B. Bridges, Ionized gas lasers, in Handbook of Laser Science and Technol. Sec. 2, Vol. II. edited by M. J. Weber Chemical Rubber Corp., Boca Raton, 1982 . 29 M. A. Dunn and J. N. Ross, Prog. Quantum Electron. 4, 233 1976 . 30 F. Irons and N. J. Peacock, J. Phys. B 7, 1109 1974 . 31 R. J. Dewhurst, D. Jacoby, G. J. Pert, and S. A. Ramsden, Phys. Rev. Lett. 37, 1265 1976 . 32 Y. A. Kononov, K. N. Koshelev, Yu. A. Levykin, Yu. V. Sidelnikov, and S. S. Churilov, Sov. J. Quantum Electron. 6, 308 1976 . 33 V. A. Bhagavatula and B. Yarkobi, Opt. Commun. 24, 331 1978 . 34 R. H. Dixon and R. C. Elton, Phys. Rev. Lett. 38, 1072 1977 . 35 M. H. Key, C. L. S. Lewis, and M. J. Lamb, Opt. Commun. 28, 331 1979 . 36 P. Jaegle, G. Jamelot, A. Carillon, and C. Wehenkel, Jpn. J. Appl. Phys., Part 1 17, 483 1978 . 37 D. Jacoby, G. J. Pert, S. A. Ramsden, L. D. Shorrock, and G. J. Tallents, Opt. Commun. 37, 193 1981 ; D. Jacoby, G. J. Pert, L. D. Shorrock, and G. J. Tallents, J. Phys. B 15, 3557 1982 . 38 D. L. Matthews et al., Phys. Rev. Lett. 54, 110 1985 ; M. D. Rosen, P. L. Hagelstein, D. L. Matthews, E. M. Campbell, A. U. Hazi, B. L. Whitten, B. MacGowan, R. E. Turner, and R. W. Lee, ibid. 54, 106 1985 . 39 S. Suckewer, C. H. Skinner, H. Milchberg, C. Keane, and D. Voorhees, Phys. Rev. Lett. 55, 1753 1985 ; S. Suckewer, C. H. Skinner, D. Kim, E. Valeo, D. Voorhees, and A. Wouters, ibid. 57, 1004 1986 . 40 T. N. Lee, E. A. McLean, and R. C. Elton, Phys. Rev. Lett. 59, 1185 1987 . 41 C. Chenais-Popovics et al., Phys. Rev. Lett. 59, 2161 1987 . 42 B. J. MacGowan et al., J. Appl. Phys. 61, 5243 1987 . 43 B. J. MacGowan, S. Maxon, P. L. Hagelstein, C. J. Keane, R. A. London, D. L. Matthews, M. D. Rosen, J. H. Scoeld, and D. A. Whelan, Phys. Rev. Lett. 59, 2157 1987 . 44 B. J. MacGowan et al., Phys. Fluids B 4, 2326 1992 . 45 A. Carillon et al., Phys. Rev. Lett. 68, 2917 1992 . 46 J. A. Koch et al., Phys. Rev. Lett. 68, 3291 1992 . 47 B. Rus et al., in Ref. 4, p. 293. 48 C. H. Skinner, D. S. Dicicco, D. Kim, R. J. Rosser, S. Suckewer, A. P. Gupta, and J. G. Hirshberg, J. Microsc. 159, 51 1990 . 49 L. B. DaSilva et al., Science 258, 269 1992 . 50 J. E. Trebes et al., Science 238, 517 1987 . 51 P. Jaegle et al., J. Appl. Phys. 81, 2406 1997 ; G. Jamelot et al., in Ref. 1, p. 124. 52 R. W. Waynant and R. C. Elton, Proc. IEEE 64, 1059 1976 . 53 P. L. Hagelstein, Plasma Phys. 25, 1345 1983 ; Atomic Physics 9, edited by R. S. Van Dyck, Jr. and E. N. Fortson World Scientic, Singapore, 1984 , p. 382. 54 C. J. Keane, Proc. SPIE 1551, 2 1991 . 55 H. C. Kapteyn, L. B. Da Silva and R. W. Falcone, Proc. IEEE 80, 342 1992 . 56 D. L. Matthews, Nucl. Instrum. Methods Phys. Res. B 98, 91 1995 . 57 C. H. Skinner, Phys. Fluids B 3, 2420 1991 . 58 R. C. Elton, X-Ray Lasers Academic, Boston, 1990 . 59 Ultrashort Wavelength Lasers, edited by S. Suckewer, Proc. SPIE Society Photo-optical Engineers, Bellingham, WA, 1991 , Vol. 1551. 60 X-Ray Lasers, edited by E. E. Fill Institute of Physics Press, Bristol, 1992 . 61 Ultrashort Wavelength Lasers II, edited by S. Suckewer, Proc. SPIE Society Photo-optical Engineerings, Bellingham, WA, 1993 , Vol. 2012. 62 S. Backus, C. G. Durfee III, M. Murnane, and H. C. Kapteyn, Rev. Sci. Instrum. 69, 1207 1998 . 63 C. P. J. Barty et al., in Ref. 2, p. 282. 64 D. P. Umstadter, C. Barty, M. Perry, and G. A. Mourou, Opt. Photonic News 9, 41 1998 . 65 Y. Kato, in Ref. 2, p. 274. 66 J. J. Rocca, M. C. Marconi, B. T. Szapiro, and J. Meyer, Proc. SPIE 1551, 275 1991 . 67 R. C. Elton, J. D. Shipman, Jr., and F. C. Young, Conference Record Abstracts 1990 IEEE International Conference on Plasma Science 1990, p. 108.
using traveling wave excitation. The laser output pulse energy of the line of Ni-like Pd was reported to be approximately 10 J for 7 J of laser pump energy.
ACKNOWLEDGMENTS
The author thanks J. Dunn, K. Frank, S. Harris, H. Kapteyn, R. London, K. Midorikawa, M. Murnane, P. Nickles, A. Osterheld, V. Shlyaptsev, and S. Suckewer for supplying the gures that illustrate their work. The useful comments of J. Chilla, R. Elton, C. S. Menoni, S. Suckewer, V. Shlyaptsev, and F. Tomasel are appreciated. In his own work in soft x-ray lasers the author acknowledges the collaboration of B. Benware, D. Beethe, D. Burt, J. Chilla, D. Clark, O. Cortazar, K. Floyd, M. Frati, J. Filevich, J. Gonzalez, D. Hartshorn, K. Kanizay, C. Macchietto, H. Mancini, M. Marconi, O. Martinez, C. Moreno, K. Richardson, V. Shlyaptsev, B. Szapiro, F. Tomasel, and A. Vinogradov. The assistance of B. Hutchenson with the mechanics of the manuscript is appreciated. The author also acknowledges the support of the National Science Foundation and the Department of Energy.
Soft X-ray Lasers and Applications II, edited by J. J. Rocca and L. B. Da Silva, Proc. SPIE Society of Photooptical Engineers, Bellingham, WA, 1997 , Vol. 3156. 2 X-Ray Lasers, edited by S. Svanberg and C. G. Wahlstrom Institute of Physics, University of Reading, Berkshi 1996 , Vol. 1511. 3 Soft X-Ray Lasers and Applications, edited by J. J. Rocca and P. L. Hagelstein, Proc. SPIE Society of Photooptical Engineers, Bellingham, WA, 1995 , Vol. 2520. 4 X-Ray Lasers, edited by D. C. Eder and D. L. Matthews AIP, New York, 1994 . 5 A. LHuillier, in Ref. 2, P. 444 and references therein. 6 Z. Chang, A. Rundquist, H. Wang, M. M. Murnane, and H. C. Kapteyn, Phys. Rev. Lett. 79, 2967 1997 . 7 A. Rundquist, C. G. Durfee III, Z. Chang, C. Herne, S. Backus, M. Murnane, and H. C. Kapteyn, Science 280, 1412 1998 . 8 T. Ditmire, J. K. Crane, H. Nyugen, L. B. Da Silva, and M. D. Perry, Phys. Rev. A 51, R902 1995 . 9 D. T. Attwood, K. Halbach, and N. J. Kim, Science 228, 1264 1985 . 10 D. T. Attwood, G. Sommargren, R. Beguiristain, K. Nguyen, J. Boker, N. Ceglio, K. Jackson, M. Koike, and J. Underwood, Appl. Opt. 32, 7022 1993 . 11 R. Coisson, Appl. Opt. 34, 904 1995 . 12 J. Rossback, in Ref. 2, p. 436. 13 Free-Electron Laser Challenges, edited by P. G. OShea and H. E. Bennett, Proc. SPIE Society of Photooptical Engineers, Bellingham, WA, 1997 , Vol. 2988. 14 P. L. Shkolnikov and A. E. Kaplan, in Ref. 2, p. 512 and references therein. 15 L. B. Da Silva et al., Phys. Rev. Lett. 74, 3991 1995 ; A. S. Wan, T. W. Barbee Jr., R. Cauble, P. Celliers, L. B. Da Silva, J. C. Moreno, P. W. Rambo, G. F. Stone, J. E. Trebes, and F. Weber, Phys. Rev. E 55, 6293 1997 . 16 R. Cauble, L. B. Da Silva, T. W. Barbee Jr., P. Celliers, J. C. Moreno, and A. S. Wan, Phys. Rev. Lett. 74, 3816 1995 . 17 B. R. Benware, C. D. Macchietto, C. H. Moreno, and J. J. Rocca, Phys. Rev. Lett. 81, 5804 1998 . 18 C. D. Macchietto, B. R. Benware, and J. J. Rocca, Opt. Lett. 24, 1115 1999 . 19 T. H. Maiman, Nature London 187, 493 1960 ; A. Javan, W. R. Bennett, Jr., and D. R. Herriot, Phys. Rev. 6, 106 1961 . 20 G. A. Gudzenko and L. A. Shelepin, Zh. Eksp. Teor. Fiz. 45, 1445 1963 Sov. Phys. JETP 18, 998 1964 . 21 M. A. Duguay and P. M. Rentzepis, Appl. Phys. Lett. 10, 350 1967 . 22 A. G. Molchanov, Sov. Phys. Usp. 15, 124 1972 . 23 R. C. Elton, Appl. Opt. 14, 97 1975 . 24 A. N. Zherikin, K. N. Koshelev, and V. S. Letokhov, Sov. J. Quantum Electron. 6, 82 1976 .
1
Rev. Sci. Instrum., Vol. 70, No. 10, October 1999 J. J. Rocca, O. D. Cortazar, B. Szapiro, K. Floyd, and F. G. Tomasel, Phys. Rev. E 47, 1299 1993 . 69 J. J. Rocca, O. D. Cortazar, F. G. Tomasel, and B. T. Szapiro, Phys. Rev. E 48, R2378 1993 . 70 J. J. Rocca, M. C. Marconi, J. L. A. Chilla, D. P. Clark, F. G. Tomasel, and V. N. Shlyaptsev, IEEE J. Sel. Top. Quantum Electron. 1, 945 1995 . 71 H. Hirose, T. Hara, K. Ando, F. Negishi, and Y. Aoyagi, Jpn. J. Appl. Phys., Part 2 32, L1538 1993 ; T. Hara, K. Ando, F. Negishi-Tsuboi, Y. Aoyagi, H. Hirose, and Y. Yashiro, Proc. SPIE 1627, 377 1992 . 72 Y. Nagata, K. Midorikawa, S. Kubodera, M. Obara, H. Tashiro, and K. Tokoda, Phys. Rev. Lett. 71, 3774 1993 . 73 J. Goodberlet, S. Basu, M. H. Nuendel, S. Kaushik, T. Savas, M. Fleury, and P. L. Hagelstein, J. Opt. Soc. Am. B 12, 980 1995 . 74 D. Korobkin, C. H. Nam, S. Suckewer, and A. Golstov, Phys. Rev. Lett. 77, 5206 1996 . 75 B. E. Lemoff, G. Y. Yin, C. L. Gordon III, C. P. J. Barty, and S. E. Harris, Phys. Rev. Lett. 74, 1574 1995 ; J. Opt. Soc. Am. B 13, 180 1996 . 76 P. V. Nickles, V. N. Shlyaptsev, M. Kalashnikov, M. Schnurer, I. Will, and W. Sandner, Phys. Rev. Lett. 78, 2748 1997 . 77 J. Dunn, A. L. Osterheld, R. Shepherd, W. E. White, V. N. Shlyaptsev, and R. E. Stewards, Phys. Rev. Lett. 80, 2825 1998 . 78 J. Dunn, A. L. Osterheld, R. Shepherd, W. E. White, V. N. Shlyaptsev, A. Bullock, and R. E. Stewart, in Ref. 1, p. 114. 79 A. R. Parag, F. Lowenthal, R. Tommasini, and J. E. Balmer, Appl. Phys. B: Lasers Opt. 66, 561 1998 . 80 D. Korobkin, A. Goltsov, A. Morozov, and S. Suckewer, Phys. Rev. Lett. 81, 1607 1998 . 81 J. J. Rocca, V. Shlyaptsev, F. G. Tomasel, O. D. Cortazar, D. Hartshorn, and J. L. A. Chilla, Phys. Rev. Lett. 73, 2192 1994 . 82 J. J. Rocca, F. G. Tomasel, M. C. Marconi, V. N. Shlyaptsev, J. L. A. Chilla, B. T. Szapiro, and G. Guidice, Phys. Plasmas 2, 2547 1995 . 83 J. J. Rocca, D. P. Clark, J. L. A. Chilla, and V. N. Shlyaptsev, Phys. Rev. Lett. 77, 1476 1996 . 84 Yu. A. Uspenskii et al., Opt. Lett. 23, 771 1998 . 85 T. W. Barbee, Jr., J. C. Rife, W. R. Hunter, M. P. Kowalski, R. G. Cruddace, and J. F. Seely, Appl. Opt. 32, 4852 1993 . 86 A. D. Siegman, Lasers University Science Books, Mill Valley, CA, 1986 , p. 301. 87 G. J. Linford, E. R. Peresini, W. R. Soor, and M. L. Spaeth, Appl. Opt. 13, 379 1974 . 88 For typical gains in visible gas lasers, see for example, W. T. Silfvast, Laser Fundamentals Cambridge University Press, Cambridge, 1996 . 89 K. Kato et al., Proc. SPIE 3156, 2 1997 . 90 V. A. Chirkov, Sov. J. Quantum Electron. 14, 1497 1984 . 91 R. A. London and M. D. Rosen, Phys. Fluids 29, 3813 1986 . 92 R. A. London, Phys. Fluids 31, 184 1988 . 93 E. E. Fill, Opt. Commun. 67, 441 1988 . 94 P. B. Holden and B. Rus, Opt. Commun. 119, 424 1995 . 95 J. L. A. Chilla and J. J. Rocca, J. Opt. Soc. Am. B 13, 2841 1996 . 96 E. E. Fill, J. Opt. Soc. Am. B 14, 1505 1997 . 97 C. H. Moreno, M. C. Marconi, V. N. Shlyaptsev, B. R. Benware, C. D. Macchietto, J. L. A. Chilla, J. J. Rocca, and A. Osterheld, Phys. Rev. A 58, 1509 1998 . 98 J. C. Moreno, J. Nilsen, Y. Li, P. Lu, and E. E. Fill, Opt. Lett. 21, 585 1996 . 99 J. Nilsen, J. C. Moreno, L. B. DaSilva, and T. W. Barbee, Jr., Phys. Rev. A 55, 827 1997 . 100 J. Zhang et al., Phys. Rev. A 54, R4653 1996 . 101 J. C. Moreno, J. Nilsen, Y. L. Li, and E. E. Fill, Opt. Lett. 21, 866 1996 . 102 J. Nilsen, J. C. Moreno, L. B. DaSilva, and T. W. Barbee, Jr., Phys. Rev. A 55, 827 1997 . 103 I. T. Boehly et al., Phys. Rev. A 42, 6962 1990 . 104 J. Nilsen, B. J. MacGowan, L. B. Da Silva, and J. C. Moreno, Phys. Rev. A 48, 4682 1993 . 105 J. Nilsen and J. C. Moreno, Phys. Rev. Lett. 74, 3376 1995 . 106 G. Jamelot et al., Proc. SPIE 2520, 2 1995 . 107 G. F. Cairns, et al., Opt. Commun. 124, 777 1996 . 108 Y. Li, G. Pretzler, and E. E. Fill, Phys. Rev. A 52, R3433 1995 ; E. F. Fill, Y. Li, D. Schlogl, J. Steingruber, and J. Nilsen, Proc. SPIE 2520, 134 1995 . 109 H. Daido, R. Kodama, K. Murai, G. Yuan, M. Takagi, Y. Kato, I. W. Choi, and C. H. Nam, Opt. Lett. 20, 61 1995 .
68 110
X-ray lasers
3825
B. Rus, A. Carillon, P. Dhez, P. Jaegle, G. Jamelot, A. Klisnick, M. Nantel, and P. Zeitoun, Phys. Rev. A 55, 3858 1997 . 111 H. Daido, S. Ninomiya, T. Imani, B. Kodama, M. Takagi, Y. Kato, K. Murai, J. Zhang, Y. You, and Y. Gu, Opt. Lett. 21, 958 1996 . 112 E. E. Fill, Y. Li, D. Schlogl, J. Steingruber, and J. Nilsen, Opt. Lett. 20, 374 1995 . 113 G. J. Tallents, Proc. SPIE 3156, 30 1997 . 114 J. Zhang et al., Phys. Rev. Lett. 78, 3856 1997 . 115 J. Zhang et al., Science 276, 1097 1997 . 116 J. Zhang et al., Phys. Rev. A 54, R4653 1996 . 117 Y. L. Li, G. Pretzler, P. X. Lu, E. E. Fill, and J. Nilsen, Phys. Plasmas 4, 479 1997 . 118 F. G. Tomasel, V. N. Shlyaptsev, and J. J. Rocca, Phys. Rev. A 54, 2474 1996 . 119 C. G. Durfee III, J. Lynch, and H. M. Milchberg, Phys. Rev. E 51, 2368 1995 . 120 H. M. Milchberg, C. G. Durfee III, and J. Lynch, J. Opt. Soc. Am. B 12, 731 1995 . 121 C. G. Durfee III, J. Lynch, and H. M. Milchberg, Opt. Lett. 19, 1937 1994 . 122 Y. Ehrlich, C. Cohen, A. Zigler, J. Krall, P. Sprangle, and E. Erasay, Phys. Rev. Lett. 77, 4186 1996 . 123 J. J. Rocca, F. G. Tomasel, M. C. Marconi, J. L. A. Chilla, C. H. Moreno, B. R. Benware, V. N. Shlyaptsev, J. J. Gonzales, and C. D. Macchietto, Proc. SPIE 3156, 164 1997 . 124 F. G. Tomasel, J. J. Rocca, V. N. Shlyaptsev, and C. D. Macchietto, Phys. Rev. A 55, 1437 1997 . 125 Y. Li, P. Lu, G. Pretzler, and E. E. Fill, Opt. Commun. 133, 196 1997 . 126 D. J. Fields et al., Phys. Rev. A 46, 1606 1992 . 127 G. J. Pert, J. Phys. B 9, 3301 1976 . 128 B. L. Whitten, R. A. London, and R. S. Walling, J. Opt. Soc. Am. B 5, 2537 1988 . 129 V. N. Shlyaptsev, J. J. Rocca, and A. L. Osterheld, Proc. SPIE 2520, 265 1995 . 130 B. L. Whitten et al., Interim Report of the J 0 Task Force, Lawrence, Livermore Lab. UCID-21152 1987 . 131 H. R. Griem, Phys. Rev. A 33, 3580 1986 . 132 J. P. Apruzese, J. Davis, M. Blaha, P. C. Kepple, and V. L. Jacobs, Phys. Rev. Lett. 55, 1877 1985 . 133 H. Fiedorowicz, A. Bartnik, Y. Li, P. Lu, and E. Fill, Phys. Rev. Lett. 76, 415 1996 . 134 S. Maxon, P. Hagelstein, J. Scoeld, and Y. Lee, J. Appl. Phys. 59, 293 1986 . 135 S. Maxon, P. Hagelstein, B. MacGowan, R. London, M. Rosen, J. Scoeld, S. Dalhed, and M. Chen, Phys. Rev. A 37, 2227 1988 . 136 W. H. Goldstein, J. Oreg, A. Zigler, A. Bar-Shalom, and M. Klapisch, Phys. Rev. A 38, 1797 1988 . 137 B. J. MacGowan, S. Maxon, C. J. Keane, R. A. London, D. L. Matthews, and D. A. Whelan, J. Opt. Soc. Am. 5, 1858 1988 . 138 P. L. Hagelstein, Proc. OSA Meeting on Short Wavelength Coherent Radiation: Generation and Applications, edited by R. W. Falcone and J. Korz, 1988 p. 28; Proc. SPIE 1551, 254 1991 . 139 B. J. MacGowan et al., Phys. Rev. Lett. 65, 420 1990 . 140 B. J. MacGowan et al., Phys. Rev. Lett. 65, 2374 1990 . 141 H. Daido et al., Phys. Rev. Lett. 75, 1074 1995 . 142 H. Daido et al., Opt. Lett. 21, 958 1996 . 143 J. Zhang et al., Proc. SPIE 3156, 53 1997 . 144 R. Tommasini, F. Lowenthal, and J. E. Balmer, in X-ray Lasers 1998, Proceedings of the 6th International Conference on X-ray Lasers, edited by Y. Kato, H. Takuma, and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 ; R. Tommasini, F. Lowenthal, and J. E. Balmer, Phys. Rev. A 59, 1577 1999 . 145 V. N. Shlyaptsev, Ph.D. thesis, P. N. Lebedev, Physical Institute, 1987, Preprint FIAN #40 1985 . 146 Yu. A. Afanasiev and V. N. Shlyaptsev, Sov. J. Quantum Electron. 19, 1606 1989 . 147 V. N. Shlyaptsev, P. V. Nickles, T. Schlegel, M. R. Kalashnikov, and A. L. Osterheld, Proc. SPIE 2012, 111 1993 . 148 V. N. Shlyaptsev, J. J. Rocca, M. P. Kalashnikov, P. V. Nickles, W. Sandner, A. L. Osterheld, J. Dunn, and D. C. Eder, Proc. SPIE 3156, 93 1997 . 149 C. D. Decker and R. A. London, Proc. SPIE 3156, 94 1997 . 150 J. Nilsen, Proc. SPIE 3156, 86 1997 . 151 C. Steden and H. J. Kunze, Phys. Lett. A 151, 534 1990 .
3826
152

196
J. J. Rocca N. Qi and M. Krishnan, Phys. Rev. Lett. 59, 2051 1987 ; Phys. Rev. A 39, 4651 1989 . 197 R. H. Dixon and R. C. Elton, J. Opt. Soc. Am. B 1, 232 1984 . 198 R. C. Elton, in Ref. 58, pp. 131142 and references therein. 199 B. N. Chichkov and E. E. Fill, Phys. Rev. A 42, 599 1990 . 200 J. Nilsen, J. Quant. Spectrosc. Radiat. Transf. 47, 171 1991 . 201 V. Yu. Politov, P. A. Loboda, V. A. Lykov, and J. Nilsen, Opt. Commun. 108, 283 1994 . 202 P. Belerdorfer, S. R. Elliot, and J. Nilsen, Phys. Rev. A 49, 3123 1994 . 203 S. J. Stephanakis et al., IEEE Trans. Plasma Sci. 16, 472 1988 . 204 C. Deeney, T. Nash, R. R. Prasad, and J. P. Apruzese, Appl. Phys. Lett. 58, 1021 1989 . 205 J. L. Porter et al., Phys. Rev. Lett. 68, 796 1992 . 206 E. E. Fill, C. Bergmann, A. Lyras, and T. Schlegel, in Ref. 60, p. 147. 207 K. J. Ilcisin, F. Aumayr, J. L. Schwob, and S. Suckewer, J. Opt. Soc. Am. B 11, 1436 1994 . 208 N. Qi, D. A. Hammer, D. H. Kalantar, and K. C. Mittal, Phys. Rev. A 47, 2253 1993 . 209 A. V. Vinogradov and I. I. Sobelman, Sov. Phys. JETP 36, 1115 1973 . 210 L. P. Presnyakov and A. D. Ulantsev, Sov. J. Quantum Electron. 4, 1320 1975 . 211 M. D. Scully, W. H. Louisell, and W. B. McKnight, Opt. Commun. 9, 246 1973 . 212 P. L. Hagelstein, Proc. SPIE 3156, 230 1997 . 213 W. Tighe et al., Rev. Sci. Instrum. 59, 2235 1998 . 214 T. S. Luk, A. McPherson, G. Gibson, K. Boyer, and C. K. Rodhes, Opt. Lett. 14, 1113 1989 . 215 S. Watanabe, A. Endoh, M. Watanabe, and N. Sarukura, in Short Wavelength Coherent Radiation: Generation and Applications, edited by R. W. Falcone and J. Kirz OSA, Washington, 1998 . 216 J. Dunn, J. Nilsen, A. L. Osterheld, Y. Li, and V. N. Shlyaptsev, Opt. Lett. 24, 101 1999 . 217 D. Strickland and G. Mourou, Opt. Commun. 56, 219 1985 . 218 E. B. Treacy, IEEE J. Quantum Electron. QE-5, 454 1969 . 219 O. E. Martinez, IEEE J. Quantum Electron. QE-23, 1385 1987 , QE-23, 59 1987 . 220 D. E. Spence, P. N. Kean, and W. Sibbett, Opt. Lett. 16, 42 1991 . 221 L. Spinelli, B. Couilland, N. Goldblat, and D. K. Negus, in Conference on Lasers and Electro-Optics, 1991, Vol. 10, paper CPDP7, OSA Tech. Digest Series Optical Society of America, Washington D.C., 1991 . 222 See Refs. 57, 65, 69, and 70 in Backus et al., Ref. 62 and Ref. 7 in D. P. Umstadter et al., Ref. 64. 223 J. P. Zhou, C. P. Hung, C. Shi, H. C. Kapteyn, and M. M. Murnane, Opt. Lett. 19, 126 1994 ; J. V. Rudd, G. Korn, S. Kane, J. Squire, and G. Mourou, ibid. 18, 2044 1993 ; B. E. Lemoff and C. P. J. Barty, ibid. 18, 1651 1993 ; W. E. White, F. G. Patterson, R. L. Combs, D. F. Price, and R. L. Shepherd, ibid. 18, 1343 1993 . 224 H. Takuma et al., in X-ray Lasers 1998, Proceedings of the 6th International Conference on X-Ray Lasers, edited by Y. Kato, H. Takuma and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 . 225 P. V. Nickles, M. Schnurer, M. P. Kalashnikov, I. Will, W. Sander, and V. N. Shlyaptsev, Proc. SPIE 2520, 373 1995 . 226 P. V. Nickles, V. N. Shlyaptsev, M. Schnurer, M. P. Kalachnikov, T. Schlegel, and W. Sandner, Opt. Commun. 142, 257 1997 . 227 W. Hartmann, H. Bauer, J. Christiansen, K. Frank, H. Kuhn, M. Stetter, R. Trotz, and T. Wagner, Appl. Phys. Lett. 58, 2619 1991 . 228 J. J. Rocca, D. C. Beetle, and M. C. Marconi, Opt. Lett. 13, 565 1988 . 229 H. Conrads, Z. Phys. 444, 200 1967 . 230 P. Bogen, H. Conrads, G. Gatti, and W. Kohlhaas, J. Opt. Soc. Am. 58, 203 1968 . 231 R. A. McCorkle, Appl. Phys. A: Solids Surf. 26, 261 1981 . 232 S. M. Zakharov, A. A. Kolomenskii, S. Pikuz, and A. I. Samokhin, Sov. Tech. Phys. Lett. 6, 486 1980 . 233 M. C. Marconi and J. J. Rocca, Appl. Phys. Lett. 54, 2180 1989 . 234 J. J. Rocca, M. C. Marconi, and F. G. Tomasel, IEEE J. Quantum Electron. 29, 182 1993 . 235 F. G. Tomasel, J. J. Rocca, O. D. Cortazar, B. T. Szapiro, and R. W. Lee, Phys. Rev. E 47, 3590 1993 . 236 C. A. Morgan, H. R. Griemond, and R. C. Elton, Phys. Rev. E 49, 2282 1994 . 237 F. G. Tomasel, J. J. Rocca, and V. N. Shlyaptsev, IEEE Trans. Plasma Sci. 24, 49 1996 . 238 T. Hosakai, M. Nakajima, T. Aoki, M. Ogawa, and K. Horioka, Jpn. J. Appl. Phys., Part 1 36, 2327 1997 ; H. Bender III, S. E. Grantham, V. N.
H. J. Shin, D. E. Kim, and T. N. Lee, Phys. Rev. E 50, 1376 1994 ; R. Dussart, W. Rosenfeld, N. Richard, D. Hong, C. Cachoncinlle, C. Fleurier, and J. M. Pouvesly, in X-ray Lasers 1998, Proceedings 6th International Conference on X-ray Lasers, edited by Y. Kato, H. Takuma, and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 . 153 T. Wagner, E. Eberl, K. Frank, W. Hartmann, D. H. H. Hoffmann, and R. Thotz, Phys. Rev. Lett. 76, 3124 1996 . 154 S. Glender and H. J. Kunze, Phys. Rev. E 49, 1586 1994 . 155 T. Boss, W. Neff, T. Boboc, F. Weigand, R. Bischoff, and H. Langhoff, J. Phys. D 31, 2472 1998 . 156 D. R. Bates and A. Dalgarno, in Atomic and Molecular Processes, edited by D. R. Bates Academic, New York, 1962 , p. 245. 157 F. V. Bunkin, V. I. Der Zhiev, and S. I. Yakovlenko, Sov. J. Quantum Electron. 11, 981 1981 . 158 See, for example, Ref. 88, p. 256. 159 G. J. Pert, J. Phys. B 9, 3301 1976 ; A. K. Dave and G. J. Pert, ibid. 18, 1027 1985 . 160 J. F. Seely, C. W. Brown, U. Feldman, M. Richardson, B. Yaakobi, and W. E. Behring, Opt. Commun. 54, 289 1985 . 161 M. Grande et al., Opt. Commun. 74, 309 1990 . 162 B. Boswell, D. Shavarts, T. Boehly, and B. Yaakobi, Phys. Fluids B 2, 436 1990 . 163 H. Azumma et al., Opt. Lett. 15, 1011 1990 . 164 G. Jamelot, A. Klisnick, A. Carillon, H. Guennov, A. Sureau, and P. Jaegle, J. Phys. B 18, 4647 1998 . 165 P. Jaegle, G. Jamelot, A. Carrillon, A. Klisnik, A. Sureau, and H. Guennov, J. Opt. Soc. Am. B 4, 563 1987 ; A. Carillon et al., J. Phys. B 23, 147 1990 . 166 D. Kim, C. H. Skinner, A. Wouters, E. Valed, D. Voorhes, and S. Suckewer, J. Opt. Soc. Am. B 6, 115 1989 . 167 H. Milchberg, C. H. Skinner, S. Suckewer, and D. Voorhees, Appl. Phys. Lett. 47, 1151 1985 . 168 S. Suckewer and H. Fishman, J. Appl. Phys. 51, 1922 1980 . 169 C. J. Keane and S. Suckewer, J. Opt. Soc. Am. B 8, 201 1991 . 170 J. Steingruber, S. S. Chen, and E. E. Fill, in Ref. 60, p. 115. 171 G. J. Pert, in Ref. 4, p. 49. 172 W. W. Jones and A. W. Ali, Appl. Phys. Lett. 26, 450 1975 . 173 J. Peyraud and N. Peyraud, J. Appl. Phys. 43, 2993 1972 . 174 P. B. Corkum, N. H. Burnett, and F. Brunel, Phys. Rev. Lett. 62, 1259 1989 . 175 B. N. Chichkov and B. Wellegehausen, in X-ray Lasers 1998, Proceedings of the 6th International Conference on X-ray Lasers, edited by Y. Kato, H. Takuma and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 . 176 N. H. Burnett and P. B. Corkum, J. Opt. Soc. Am. B 6, 1195 1989 . 177 P. E. Amendt and D. C. Eder, Phys. Rev. Lett. 66, 2589 1991 ; D. C. Eder, P. Amendt, and S. C. Wilks, Phys. Rev. A 45, 6761 1992 . 178 D. C. Eder et al., Phys. Plasmas 1, 1744 1994 . 179 P. Polonsky, C. O. Park, K. Krushelnick, and S. Suckewer, Proc. SPIE 2012, 61 1993 . 180 H. C. Kapteyn, Appl. Opt. 31, 4931 1992 . 181 G. L. Strobel, D. C. Eder, R. A. London, M. D. Rosen, R. W. Falcone, and S. P. Gordon, Proc. SPIE 1860, 157 1993 . 182 R. C. Elton, Appl. Opt. 14, 2243 1975 . 183 W. T. Silfvast, J. J. Macklin, and D. R. Wood II, Opt. Lett. 8, 551 1983 ; H. Lundbergt, J. J. Macklin, W. T. Silfvast, and O. R. Wood, Appl. Phys. Lett. 45, 335 1984 . 184 E. J. McGuire, Phys. Rev. Lett. 35, 844 1975 . 185 H. C. Kapteyn, R. W. Lee, and R. W. Falcone, Phys. Rev. Lett. 57, 2939 1986 . 186 M. H. Sher, S. J. Benerofe, J. F. Young, and S. E. Harris, J. Opt. Soc. Am. B 8, 114 1991 . 187 H. C. Kapteyn and R. W. Falcone, Phys. Rev. A 37, 2033 1988 . 188 S. J. Moon, D. C. Eder, and G. L. Strobel, in Ref. 4, p. 262. 189 D. L. Matthews, in Ref. 2 p. 32. 190 A. Vinogradov, I. Sobelman, and E. Yukov, Sov. J. Quantum Electron. 5, 59 1975 . 191 B. A. Norton and N. J. Peacok, J. Phys. B 8, 989 1975 . 192 J. Nilsen, H. Fiedorowicz, A. Bartnik, Y. L. Li, P. X. Lu, and E. E. Fill, Opt. Lett. 21, 408 1996 . 193 J. Nilsen, Phys. Rev. A 53, 4539 1996 . 194 J. Trebes and M. Krishnan, Phys. Rev. Lett. 50, 679 1983 . 195 N. Qi, H. Kilic, and M. Krishnan, Appl. Phys. Lett. 46, 471 1985 .
Rev. Sci. Instrum., Vol. 70, No. 10, October 1999 Shlyaptsev, J. J. Rocca, M. C. Richardson, and W. T. Silfvast, in X-ray Lasers 1998, Proceedings 6th International Conference on X-Ray Lasers, edited by Y. Kato, H. Takuma and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 . 239 J. J. Rocca, O. D. Cortazar, B. T. Szapiro, F. G. Tomasel, and D. Hart shorn, Proc. SPIE 2012, 67 1993 ; J. J. Rocca, F. G. Tomasel, C. A. Moreno, V. N. Shlyaptsev, M. C. Marconi, B. A. Benware, J. J. Gonzalez, J. L. A. Chilla, and C. D. Macchietto, J. Phys. Paris , Colloq. 7, C4-353 1997 . 240 V. N. Shlyaptsev, A. V. Gerusov, A. V. Vinogradov, J. J. Rocca, O. D. Cortazar, F. Tomasel, and B. Szapiro, Proc. SPIE 2012, 99 1993 ; V. N. Shlyaptsev, J. J. Rocca, P. V. Nickles, M. P. Kalashnikov, and A. L. Osterheld, in Ref. 2. 241 A. Bobrova, S. V. Bulanov, T. L. Razinkova, and P. V. Sasorov, Plasma Phys. Rep. 22, 349 1996 ; R. Nemirovsky, A. Ben-Kish, M. Shuker, and A. Ron, in X-ray Lasers, Proceedings 6th International Conference on X-ray Lasers, edited by Y. Kato, H. Takuma and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 . 242 J. J. Gonzalez, M. Frati, J. J. Rocca, and V. N. Shlyaptsev, in X-ray Lasers 1998, Proceedings 6th International Conference on X-Ray Lasers edited by Y. Kato, H. Takuma, and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 . 243 P. G. Burkhalter, G. Mehlman, F. C. Young, S. J. Stephanakis, V. E. Scherrer, and D. A. Newman, J. Phys. Paris , Colloq. 47, C-247 1986 see Ar spectra in Fig. 6 . 244 A. Hildebrand, A. Ruhrmann, S. Maurmann, and H. J. Kunze, Phys. Lett. A 221, 335 1996 . 245 B. R. Benware, C. H. Moreno, D. J. Burd, and J. J. Rocca, Opt. Lett. 22, 796 1997 . 246 C. H. Moreno, M. C. Marconi, V. N. Shlyaptsev, and J. J. Rocca, IEEE Trans. Plasma Sci. 27, 6 1999 . 247 M. C. Marconi, J. L. A. Chilla, C. H. Moreno, B. R. Benware, and J. J. Rocca, Phys. Rev. Lett. 79, 2799 1997 . 248 R. A. London, M. Strauss, and M. D. Rosen, Phys. Rev. Lett. 65, 563 1990 . 249 M. D. Feit and J. J. A. Flech, J. Opt. Soc. Am. B 7, 2048 1990 . 250 B. Rus, these de LUniversite Paris VII. Orsay, Jan., 1995, Proc. SPIE 3156, 17 1997 ; F. Albert et al., ibid. 3156, 247 1997 997. 251 F. C. Young, S. J. Stephanakis, V. E. Scherrer, B. L. Welch, G. Melimav, P. G. Burkhalter, and J. P. Apruzese, Appl. Phys. Lett. 50, 1053 1987 . 252 S. K. Kukhlevsky, L. Kozma, L. Palladino, A. Reale, F. Flora, and G. Giordano, Proc. SPIE 3156, 1 1999 . 253 M. Shuker, A. Ben-Kish, R. A. Nemirovsky, and A. Ron, in X-ray Lasers, Proceedings 6th International Conference on X-Ray Lasers, edited by Y. Kato, H. Takuma, and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 .
254
X-ray lasers
3827
M. Pockl, R. Fertner, M. Hebenstreit, T. Neger, and F. Aumayr, Proc. SPIE 2520, 379 1995 ; M. Hebenstreit, R. Fertner, T. Neger, M. Pockl, and F. Aumayr, J. Phys. D 29, 1933 1996 . 255 H. J. Kunze, K. N. Koshelev, C. Steden, D. Uskov, and H. T. Wieschebrink, Phys. Lett. A 194, 183 1994 . 256 P. B. Corkum and N. H. Burnett, Short-Wavelength Coherent Radiation: Generation and Applications, edited by R. Falcone and J. Kirz, OSA Proceedings Series, Vol. 2 OSA, Washington, D.C., 1988 . 257 B. E. Lemoff, C. P. J. Barty, and S. E. Harris, Opt. Lett. 19, 569 1994 . 258 S. M. Hooker, P. T. Epp, and G. Y. Yin, J. Opt. Soc. Am. B 14, 2735 1997 . 259 S. M. Hooker and S. E. Harris, Opt. Lett. 20, 1994 1995 . 260 A. Goltsov, D. Korobkin, A. Morozov, and S. Suckewer, Proceedings 1998 International Congress on Plasma Physics and 25th EPS Conference on Controlled Fusion and Plasma Physics, Prague, 1998. 261 Y. Nagata, K. Midorikawa, S. Kubodera, M. Obara, H. Tashiro, K. Toyoda, and Y. Kato, Phys. Rev. A 51, 1415 1995 . 262 B. N. Chichkov, A. Egbert, H. Eichmann, C. Momma, S. Nolte, and B. Wellegehausen, Phys. Rev. A 52, 1629 1995 . 263 M. P. Kalachnikov et al., Phys. Rev. A 57, 4778 1998 ; P. V. Nickles et al., Proc. SPIE 3156, 80 1997 . 264 D. Kim, C. H. Skinner, G. Umesh, and S. Suckewer, Opt. Lett. 14, 665 1989 . 265 D. Kim, C. H. Skinner, and S. Suckewer, Proc. SPIE 1551, 295 1991 . 266 L. Polonsky, C. O. Park, K. Krushelnick, and S. Suckewer, Proc. SPIE 2012, 75 1993 . 267 A. Morozov, L. Polonsky, and S. Suckewer, Proc. SPIE 2520, 180 1995 . 268 T. Hara, K. Ando, Y. Aoyagi, N. Kusakabe, and H. Yashiro, Jpn. J. Appl. Phys., Part 2 28, L1010 1989 . 269 N. Yamaguchi, T. Hara, C. Fujikawa, and Y. Hisada, Jpn. J. Appl. Phys., Part 2 36, L1297 1997 . 270 S. Basu, P. L. Hagelstein, and J. G. Goodberlet, Appl. Phys. B: Photophys. Laser Chem. 57, 303 1993 . 271 P. Hagelstein, M. Muendel, J. Goodberlet, S. Basu, and S. Kaushik, Proc. SPIE 1551, 88 1991 . 272 A. R. Prag, F. Loewenthal, and J. E. Balmer, Phys. Rev. A 54, 4585 1996 . 273 P. Hagelstein private communication . 274 J. J. Rocca, C. H. Moreno, M. C. Marconi, and K. Kanizay, Opt. Lett. 24, 420 1999 ; J. J. Rocca et al. in X-ray Lasers, edited by Y. Kato, H. Takuma, and H. Daido Institute of Physics, University of Berkshire, Reading, 1999 .

J. J. Rocca - Table-Top Soft X-Ray Lasers

Uploaded by

Copyright:

Available Formats

J. J. Rocca - Table-Top Soft X-Ray Lasers

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

J. J. Rocca - Table-Top Soft X-Ray Lasers

Uploaded by

Copyright:

Available Formats

REVIEW OF SCIENTIFIC INSTRUMENTS

VOLUME 70, NUMBER 10

REVIEW ARTICLE Table-top soft x-ray lasers

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

where n ec is the critical density n ec m ec 2 e2

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

V. SPECIFIC LASER SYSTEMS A. Capillary discharge lasers

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

1. Collisionally excited Pd-like Xe laser in an opticaleld-ionized plasma

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

to shorter wavelengths utilizing Be-like ions.259

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

Rev. Sci. Instrum., Vol. 70, No. 10, October 1999

You might also like