Ji-&

*p I@ l-

Nuclear

Instruments

and Methods

in Physics

Research A 398

( 1997)

55-64

__

NUCLEAR INSTRUMENTS 8 METHOOS IN PHYSICS RESEARCH

SectIonA

ELSEYIER

Atomic physics using short-wavelength

Sune Svanberg, Anne LHuillier

coherent radiation

and Claes-G&-an Wahlstrijm

Abstract Laser spectroscopy of free atoms has been pursued for 25 years starting in the visible region and with a constant thrust towards shorter wavelengths. A review of generation schemes and atomic-physics experiments in the shor-wavelength region is given. Processes in ions and inner shells occur on a very fast timescale setting special demands on the sources. Using picosecond laser technology efficient frequency conversion and an adequate time resolution is achievable, but the spectral resolution is often impaired with such sources. Methods to achieve temporal as well as spectral resolution in the short-wavelength region are discussed. The use of high-order harmonics as tunable short-pulse radiation is described more in detail. Future possibilities for laser spectroscopy and atomic physics at short wavelengths are finally discussed.

1. Introduction

High-resolution laser spectroscopy in the time and frequency domains became possible with the development of tunable laser sources, and the introduction of new spectroscopic techniques (see, e.g. Refs. [ 1.21). The motivation for these studies, which were first performed in the visible region, was to obtain new information about the structural and dynamic properties of free atoms (fine- and hyperfine structures, isotopic shifts. Zeeman- and Stark effects, natural radiative lifetimes. oscillator strengths and photoionization cross sections), but also to study more fundamental aspects such as to determine the Rydberg constant. Lamb shifts and to develop time/length standards. To study light elements such as hydrogen and helium, ionic systems, and inner-shell processes, the spectroscopic techniques must be extended to the short-wavelength region. Synchrotron radiation has for a long time constituted a general-purpose source of short-wavelength radiation. Another possibility which has the advantage of being a table top source is to generate a short-wavelength continuum from a laser-produced plasma [3]. However, in both cases the radiation is incoherent and spectrally broadbanded. The need to monochromatize the radiation results in a limited resolution and low excitation efficiencies, which makes difficult the studies of free atoms, exhibiting sharp absorptive structures. The development of thirdgeneration synchrotron sources, with much higher photon fluxes, will probably allow to reach a better spectral resolution, through highly resolving monochromators [4]. While synchrotron sources and laser-produced plasmas give access to the short-wavelength region, atomic-physics problems frequently require more narrow-band radiation
0168.9002/97/$17.00 Copyright 8 1997 Elsevier Science B.V. All

and/or higher peak powers. Recently, there has been a significant progress in short-wavelength coherent sources. In this contribution, we review the latest development in non-accelerator based short-wavelength laser light generation and its applications in atomic physics. A review of the techniques used up to a few years ago can be found in Refs. [.5,6]. The new schemes for short-wavelength freeelectron lasers, mostly based on the SASE (self-amplitied spontaneous emission) concept, are being discussed in the present volume [7,8] and will not be covered here. These schemes have a potential far beyond the performances of optical klystron sources [9.10] and could be of great interest for some of the applications discussed in the present work. The paper is organised as follows. First, we review the different coherent sources that can be used to access the short-wavelength region (<200 nm; >6 eV). In Section 3. we give examples of atomic-physics studies performed so far with these sources. Finally, we discuss future possibilities in Section 4.

2. Generation

of short-wavelength

coherent radiation

The first three methods presented below are based upon nonlinear frequency conversion schemes using tunable lasers of nanosecond pulse duration. The resulting shortwavelength pulses have a pulse energy ranging from tens of p.J up to tens of mJ. The primary laser radiation normally has a linewidth of a few GHz. and thus the linewidth of the short-wavelength radiation matches or exceeds by a factor of up to IO typical Doppler-broadened atomic transition widths. By using pulsed lasers that are injection-seeded by narrow-band continuous radiation,
rights reserved

PII

SOl68-9002(96)01227-2

56

S.

Swrrtxyc:et ctl. I Nd. htr

(typically 100 MHz) can
be

much smaller achieved.

linewidths

I? 1. Hun?Kmic generatio?
crystalline muterinls

und frequency

mi.uing in

The shortest wavelength obtainable by direct frequency doubling is 205 nm using a BBO (P-barium borate) crystal. By mixing frequency-doubled radiation with the fundamental (effectively frequency tripling) in a BBO crystal, wavelengths as short as 197 nm can be reached. The practical limit for short-wavelength generation using nonlinear processes in crystals is given by the absorption cutoff in BBO ( 189 nm). This wavelength can be reached. for example, by mixing 230 nm radiation with the fundamental of the Nd:YAG laser (1.06 pm).
1.2. Stimuluted crnti-Stokes Raman stuttering in gases

laser is set to the two-photon resonant wavelength of 212.55 nm for krypton, and a tunable photon is then subtracted as shown in Fig. I. where the attainable wavelengths for dye and titanium-sapphire lasers are also indicated [I I]. In a practical experimental set-up. the limit is set by the transmission of lithium fluoride windows at I IO nm. In order to reach shorter wavelengths. higher-order nonlinear processes are needed. Such processes require higher laser intensities. usually obtained by lasers with very short pulse durations, down to about 100 fs. This leads to a substantial Fourier broadening of the source impairing the spectral resolution available. However. the high peak power, the large spectral region available, the short pulse duration and the table top size. make high-order harmonic generation in gases a source interesting for a number of applications.
2.4. High-order harmorlic generation in pses

By stimulated vibrational anti-Stokes Raman scattering in gases at a pressure of a few bar. coherent radiation can be shifted towards shorter wavelengths. The hydrogen molecule has the largest vibrational splitting. 415.5 cm , corresponding to -0.5 eV Using tunable radiation generated by nonlinear processes in crystals (therefore above the BBO cutoff of 189 nm), in combination with high-order Raman shifting in hydrogen, wavelengths of about 160 nm can be reached.
2.3. Four-wave mi.virzg in guses

Shorter wavelengths can be reached through nonlinear processes in gases. For example, using krypton gas, tunable VUV wavelengths approaching 106 nm can be obtained by difference four-wave mixing. One tunable

High-order harmonics are becoming available as a practical source of short-wavelength radiation through the development of table-top terawatt laser systems based on chirped-pulse amplification [ 121. Using laser pumping of titanium-doped sapphire (Ti:S). systems with repetition rates as high I kHz are now becoming available. The layout of the 10 Hz terawatt laser system presently used at the Lund High-Power Laser Facility [ 131 is shown in Fig. 2. Pulses from a 100 fs mode-locked Ti:S oscillator are temporally stretched by a factor 2500 before amplification in a regenerative 14-pass unit. boosting the energy by a factor of 10. A final four--pass butterfly amplifier is used to further increase the pulse energy. After recompression to the original 100 fs pulse duration, peak powers approaching 2 TW are obtained.

465P

(b)
A LiF I cut-off: B Ti:Sapphire laser

Dye laser

Fig. 1. Scheme of tunable VUV laser radiation generation using two-photon resonant sum-difference Available wavelength regions using different tunable lasers are indicated. (From Ref. [ 111.)

frequency

mixing in hrypton

gas.

S. Sranherg

et al. I Nucl. Instr. and Met/t. in Phys. Res. A .JY8 (1997) X-63

mirig!Fl,~~~~~,
lo-' / 16 14 12 10 Wavelength (nm)

Fig. 3. Harmonics from a He pulsed jet when the focused laser intensity was approximately 10 W/cm. (From Ref. 161.)

Fig. 2. Lay-out of the Lund terawatt laser operating with chirpedpulse amplification in titanium-doped sapphire. (From Ref. [ 131.)

Overviews of high-harmonic generation and applications have been given. i.e. in Refs. [14,15]. We give here only a brief description of the main properties of this radiation.
The harmonics into a pulsed monochromator, are produced jet as of inert by focusing gases and an intense separated laser by a

shown in Fig. 3 [ 161. A spectrum showing generated odd harmonics from helium gas is presented in Fig. 4. The shortest wavelength obtained through high-order harmonic conversion is of the order of 7 nm. Different lasers from infrared to UV have been used to reach this short wavelength. The corresponding harmonic orders are the 109th for a titanium sapphire 806 nm laser [ 171. the 143rd in the case of a 1ps Nd-glass laser system ( 1053 nm) [IS] and, recently. the 37th for a KrF laser (248 nm) [19]. The emission of high-order harmonics is now relatively well understood [20,14.15]. Microscopically, it happens when an electron tunnels through the barrier due the Coulomb field combined with the laser electric field. is accelerated in the optical field, is forced back towards the nucleus when the laser field switches direction and recombines with the ion. emitting harmonics. The highestorder harmonics are obtained for the light inert gases such

as He and Ne. but the most efficient generation of (lowerorder) harmonics occurs in the heavy gases such as Kr and Xe. Ionized media have been proposed for the generation of very short wavelengths. Experimentally, however. ions have been proved to be very inefficient. especially when using low-frequency lasers [21.22]. This is attributed to the influence of free electrons which deteriorate propagation of the fundamental and generated beams. The optimization of the high-order harmonic conversion efficiency and the measurements of the spatial and temporal characteristics of the radiation have been the object of many studies [23-L%]. Typical performances for harmonic generation using a 100 fs terawatt laser are given in Table I as well as those for synchrotron undulator radiation. From the table it is clear that the harmonics are particularly attractive in terms of the radiation peak power, as well as the short pulse duration. Besides, it presents interesting possibilities for pump/probe experiments and it has the advantages of being a truly table top source. Harmonics are tunable, when the primary laser radiation is obtained from a tunable laser, such as a short-pulse, high-power dye laser or a titanium-doped sapphire (Ti:S) unit based on the chirped-pulse amplification (CPA)
Table I Estimated characteristics for radiation at lOOeV, generated as high-order harmonics of a 15Ofs Ti:sapphire laser or in an undulator at a I.5 GeV storage ring Harmonics 150mJat 1.5eV 15Ofs. 7J= lomx Undulator Max II 200 mA

Tomidal Minor

1.s ciev
Photons/pulse Repetition rate pbC.d,C Pulse width P,,,, Bandwidth lox IOHz I5 nW 0.1 ps I5 kW 0. I % 10C 500 MHz I mW 60 ps 30 mW 0.1%

Fig. 3. Experimental set-up for the study of high-order harmonica. The insert shows an arrangement for the study of ion formation. a competitive process to the harmonic generation. (From Ref. [ 161.)

scheme. However, in the first case. the maximum energy available is not higher than 25-30 eV, and in the latter case. the need for wavelength-dependent adjustments in the CPA chain can make the tuning procedure tedious. An alternative way to obtain continuously tunable radiation is to mix an intense laser at a fixed frequency, w,. with tunable radiation, y. from an optical parametric oscillator/ amplifier [29], which is pumped by a fraction of the fixed-frequency output. The two beams are mixed in the rare-gas jet and new frequency components nw, inzw, are generated. Here, 12 normally is a large number and no a small number. Tunable radiation up to 70 eV has been obtained 130.3 I]. In many spectroscopic applications the linewidths achieved with short-pulse pump lasers are too large, since the atomic levels to be interrogated do not have short lifetimes (and corresponding large Heisenberg uncertainty) to benefit from the very short laser pulses. It is then better to utilize a laser system with a longer pulse duration in order to generate harmonics with a relatively narrow bandwidth, ideally matching the Doppler width of the atomic transition under study. Such a dedicated laser system has been constructed at the High-Power Laser Facility in Lund [32.33] and is shown in Fig. 5. A distributed feed-back dye laser (DFDL) is pumped by the frequency-doubled output of an Sops Nd:YAG laser, operating at IO Hz. The dye laser output pulses, easily tunable, are amplified in two dye stages and are finally amplified in a four-pass titanium-sapphire stage, pumped by a large-frame nanosecond Nd:YAG laser. The resulting

pulses of 50 mJ energy and SO ps duration are well suited for the generation of relatively narrow-bandwidth harmonics (7-3lst) in a Xe or Kr gas jet. An application of this source will be described in the next section. An alternative approach to the frequency up-conversion techniques described above for generation of short-wavelength coherent radiation is to create population inversion in highly ionized media in order to obtain lasing directly at short wavelengths. 2.5.
Direct X-rav laser generution in plasm-bused

s0urcc.s

Soft X-ray lasing was demonstrated for the first time 12 years ago in neon-like selenium [34]. Initially, very large and low-repetition rate lasers were required for achieving X-ray lasing. The shortest wavelength achieved so far is 3.5 nm for the case of nickel-like gold [35]. Of particular interest in terms of achieving practical sources is the research performed with the new high-repetition rate, highpeak-power and moderate-pulse-energy systems [36-391, and the demonstration of direct electrical pumping of discharge-based devices [40]. However, these lasers are still in the development stage, and the number of demonstrated applications in atomic physics is very limited. Recent progress in the field of X-ray lasers is comprehensively covered in the proceedings of the last International Conferences on X-ray Lasers [41-431. At the end of this section we would like to also include some comments on the generation of broad-band,

:YAG 0 ns

200 ml

watmg
Fig. 5. Layout of a spectroscopic high-peak-power, narrow-band harmonics. (From Refs. [3?,33].) fully tunable hybrid dye/Ti:S laser system, suitable

50 ml 80~s
700-900 nm for the generation of tunable,

Fig. 6. Experimental set-up for the generation of ultra-short broad-band hard X-ray radiation. (From Ret. (481.)

synchrotron-like radiation, based on laser-produced plasmas. Although incoherent and broad-banded, such radiation of picosecond duration provides many new possibilities for pump/probe-type experiments in combination with sharp, short-wavelength sources.
2.6. Brood-band plasmas X-my genertrtion ,from laser-produced

Broad-band generation of radiation in the 100 eV region using laser-produced plasmas. produced by nanosecond laser pulses have been applied in atomic-physics research for quite some time (see e.g. Refs. 13.441). With the newly available ultrashort terawatt laser pulses, focused onto high-nuclear-charge elements, X-ray radiation, covering the whole spectrum up to MeV energies, has been obtained [45.46]. An experimental set up is shown in Fig. 6. Such radiation, which can be expected to find many applications in atomic physics, has initially been used in biological and medical imaging applications [46-491. Recently, several two-color experiments involving lasers and soft X-rays from laser-produced plasmas have been reported in the literature: one of these experiments studies laser-assisted Auger processes in xenon [50], another, laser EXAFS in SF,, 1511.

3. Laser spectroscopy

in the short-wavelength

regime

3.1. Nutuml radiutix,e 1iJetimeJ The most straightforward type of laser spectroscopy investigations at short wavelengths is the measurement of natural radiative lifetimes of excited states using timeresolved techniques. Natural lifetimes and associated transition probabilities and oscillator strengths are of interest, e.g., for testing atomic calculations and for astrophysical applications. With the availability of VUV spectra from the high-resolution Echelle spectrometer on board the Hubble Space Telescope, the need for matching atomic-physics data is rapidly increasing. Examples of such studies are lifetime determinations for As I [52], Ru II [53], and Pt II

[54], all performed in the wavelength region around 194 nm, reached by frequency mixing in crystals as described in Section 2.1. Sample ionic species were produced by forming a weakly ionized plasma using laser ablation of a powder target as described in Ref. [53]. Sequences of Rydberg states, accessible in the VUV region, are of interest for studying configuration interaction perturbations. The Bi I spectrum could be studied [55] for wavelengths down to 179 nm using stimulated anti-Stokes Raman shifting as discussed in Section 2.2, while fourwave mixing in krypton down to IX nm (Section 2.3) was used in studies of the S I 1561. Mg 1 [57]. Zn I [58], Cu I [ 1I] and Au I 1591 spectra. An experimental set up for these types of experiments is shown in Fig. 7. Examples of time-resolved spectra for sulphur are shown in Fig. 8. Free sulphur atoms were obtained by laser ablation of a PbS powder. Multi-contiguration Hartree-Fock (MCHF), relativistic Hartree-Fock (RHF) and superstructure theory (SST) approaches have been used for interpreting the data. In all the studies referred to above the characteristic exponential decays could be studied by directly recording the fluorescence light decay as a function of time using a VUV sensitive photomultiplier detector connected to a transient digitir.er. When the excitation cross sections or the available energy for the excitation are low, more efficient detection methods are necessary. For very short lifetimes, electronics cannot follow the decay and optical (pump/probe) methods must be employed. A pump/probe method combining both high detection efficiency and high time resolution was used in a demonstration experiment to measure the radiative lifetimes of the 2p and 3p P, states of helium [14,60]. The excitation radiation source was a high-order harmonic (see Section 2.4) of the ps laser described in Fig. 5. The experimental set-up used is shown in Fig. 9. To excite the 2p state (with a transition energy of 21 eV i.e. 58.4 nm), the DFDL laser was tuned to 760 nm and the 13th harmonic was selected. For the 3p case. requiring 54 nm (23 eV), the DFDL was tuned to 752 nm, the laser radiation was frequency doubled in a KDP crystal and the 7th harmonic of this UV radiation was selected. The harmonics were generated in a pulsed Kr jet and the individual components were separated with a normal-incidence spherical grating. Helium atoms were excited and the number of exited atoms was probed by photoionization using a temporally delayed pulse, chosen to be the second or third harmonic of the Nd:YAG pump laser, for the 3p or 2p state. respectively. The resulting ions were detected by a microchannel plate detector after passage though a time-of-flight spectrometer tuned to the helium mass and with a 100% collection efficiency. The spectral profile of the 13th harmonic, as obtained by recording the number of ions as a function of laser wavelength for a fixed ionization delay is given in Fig. 10. The linewidth is 0.01 nm, which is still a few tens of times the bandwidth limit (note that this relatively large bandwidth came from the DFDL laser, rather than from the

INJECTION SEEDER
VACUUM SYSTEM

Y
KDP DOUBLING CRYSTAL RETARDING PLAIX BBO CRYSTAL FOR TRIPLING

MONOCHROMATOR

l
Nd:Y AG LASER

PELLIN-BROCA PRISM

MONOCHROhtATOR

L__

FfEOLENCY

INJECTION SEEDER

E I set-up for time-resolved

RlEQUENCY TRIPLING laser spectroscopy using a four-wave mixing laser source. (From Ref. [57].)

Fig. 7. Experimental

harmonic
match

generation

process).

However,

reasonable

to the 0.0003 nm Doppler width of the helium atoms

is obtained. An experimental decay curve for the 3p state is given in Fig. I I. Detailed studies with the helium atomic density varying over three orders of magnitude showed no influence of

Fig. 9. Set-up for delayed photoionization measurements helium atoms, excited by high-harmonic radiation. (From

on Ref.

1W.)

50

100

150

200
58.41 58.42 58 43 58.44 58 45

Time (ns)
Fig. 8. Recordings of fluorescence atomic states excited by short-pulse light transients VUV radiation. for sulphur (From Ref.

Wavelength

(nm) in the

Fig. 10. Spectral profile of the 13th harmonic generated set-up described in Figs. 5 and 9. (From Ref. [60].)

S. Svanberg

et al.

I Nucl. Instr.

and Meth.

in Phw.

Res. A 398 (1997) A-64

61

Y
2 5 qfi?--

10-j -I

I I

J
7

15 MACNETICFELD

Delay (ns) Fig. 11. Experimental decay curve for the 3p P, state of He. (From Ref. [14].)

k,,,=202

nm

: : :

(mT)

resonance trapping or collisions on the measured lifetime. This shows that the XUV pulse intensities achieved with our system are adequate for general-purpose time-resolved laser spectroscopy in the XUV region.
a(63Cu)=61.7(9)MHr

3.2.

Structure

studies I
3

b(%)=4.9(7)

MHz

Studies of atomic structure are generally more demanding than lifetime studies, since special requirements on the spectral linewidth pertain. However, it is possible to use resonance or coherence techniques in conjunction with a broad-band excitation, by combining with radio-frequency induction of sublevel transitions, or by observing levelcrossing or quantum-beat phenomena following coherent excitation. We will start with a description of this second class of experiments for which the demands on the laser linewidth are more relaxed. Historically, it can be noted that the same techniques were once used [61.62] to move structural laser spectroscopy away from predominant sodium D lines in the long-wavelength part of the visible spectrum. where single-mode CW lasers were initially only available. Pulsed double-resonance techniques were used to determine the hyperfine structure of the 5p zP?,2 state of Ag reachable with pulse 206 nm radiation generated as described in Section 2. I. Radio-frequency resonances in the Paschen-Back region of the hyperfine structure were observed [63]. A 3% accuracy for the magnetic dipole interaction constant was obtained, corresponding to 0.25 MHz, which is less than 1 : IO of the laser linewidth. The level-crossing technique could also be extended to the short-wavelength region using pulsed lasers and frequency conversion as described in Section 2.1. Such an experiment was performed at 202 nm yielding precision values for the hyperfine structure constants of the 5p P3,, state of copper [64]. Experimental as well as theoretical (Breit formula) level-crossing curves are shown in Fig. 12. The third method belonging to this class is quantum-beat

2.5 MAGNETIC

5.0 FIELD

7.5 (mT)

10. i-

Fig. 12. Experimental and theoretical level-crossing curves for the 5P +,,z state of Cu. The experimental curve was obtained using pulsed. linearly polarized laser light at 202 nm. and the theoretical curve was calculated with the Breit formula using the fitted hyperfine-structure parameters. The energy-sublevel diagram is included with level-crossings indicated. (From Refs. [64,6].)

spectroscopy, where signals from coherently excited sublevels are observed in the time domain. The method was extended to the VUV region in the hyperfine-structure determination for the 7p 3P.s,2 state of Ag. requiring I85 nm excitation following stimulated Raman shifting (Section 2.2). Experimental and theoretical curves are shown in Fig. 13 [65]. The time-resolved techniques using harmonics in the XUV region as illustrated in Fig. 9 could in principle be combined with quantum-beat spectroscopy yielding data of the type shown in Fig. 13. Multiple pulse excitation [66] extending Ramsey-fringe two-pulse techniques [67] also provide high resolution. Since these techniques are independent of wavelength, it should in principle be possible to extend it to the short-wavelength region. All the techniques discussed above do not require narrow-band lasers. However, if narrow-band short-wavelength radiation is available. all the Doppler-free laserspectroscopic techniques utilized at longer wavelengths are in principle applicable. The development has here been driven by fundamental physics requirements. Thus narrow-

Fig. using

13. Experimental
pulsed. linearly

and theoretical polarized

quantum-beat curve (lower)

curves for the was obtained and

7p zP?,? state of A@. The experimental theoretical lifetime curve was calculated (From

O.ll_,_._
14.2 14.3 Photon 14.4 14.5 14.6 Energy (eV)

laser light

at IX.5 nm. and the

with fitted hyperfine-structure

parameters.

Ref. [65].)

CW radiation at 243 nm was achieved for Lamb-shift measurements in hydrogen [68.69]. At still shorter wavelengths (194 nm) minute amounts of narrow-band CW radiation was generated for atomic-clock applications using Hg ions [70]. Still shorter wavelengths require pulsed sources. with pulse duration as long as possible (in the nanosecond range) and Fourier-limited linewidths. Using tifth harmonic generation of frequency-doubled dye laser radiation. the resonance line of helium could be measured with extreme accuracy, allowing Lamb-shift determination [71.73] (Fig. 14). To get a narrow linewidth (950 Mhz at 58.4 nm). a CW seed signal was used as input in the laser amplification chain. This promising development parallels those for hydrogen Lamb-shift measurements, where pulsed amplitiers were used [73] before narrow-band CW radiation became available at 243 nm band

Fig. I.S. Photoionization cross section of krypton using the 7th harmonic of a picosecond dye laser tuned in the range S90610 nm. (From Ref. [74].)

[6X1. 3.3. Further short-M,allelenRth laser studies

of cuoms

Photoionization experiments are conveniently performed by monochromatized synchrotron radiation. However, for

broad structures discrete harmonics can be used without any need for tuning. In this way relative photoionization cross sections for Ne, Ar. Kr and Xe were measured 1741 over the range IO to I IO eV using the I I th to the 69th harmonic of a I40 fs Cr:LiSAF laser operating at 815 run. The harmonics were selected with a I m monochromator with a toroidal mirror and a plane grating. and the photoion yield was measured using a time-of-flight spectrometer with microchannel plate detector. Good agreement with synchrotron radiation data was obtained. Photoionization spectroscopy with higher resolution was performed using harmonics from a I ps transform-limited dye laser [73]. Tuning this dye laser in steps of 0.25 nm over the range 590-610 nm and selecting the 7th harmonic. generated in a Xe jet, relative photoionization cross sections of krypton between the two ionization thresholds could be determined. The results are presented in Fig. 15. High-order harmonic radiation has also been utilized for time-resolved as well as high-resolution photoelectron spectroscopy experiments 1751.

4. Outlook
Based on the experience from short-wavelength laserspectroscopy experiments performed so tar (Section 3) using the technology presently available (Section 2) it is possible to speculate on future possibilities. These depend naturally on the improvements in the source characteristics, both for conventional laser approaches and prqjected SASE FELs for the short-wavelength region (7.81. Since synchrotron work has been performed in the wavelength region of interest for a long time [76]. spectroscopy experiments requiring source performance (intensity. pulse duration. resolution, coherence), beyond what is possible with synchrotrons are of particular interest.

uvJIII,1w(lill
584 nm
I176 d etalon

markers

Fig. 14. Recording of the He resonance line at 58.4 nm using the 5th harmonic of I frequency-doubled injection-seeded pulsed dye laser system. (From Ref. [72].)

S. Svmherg

et ul.

I Nuti.

Instr.

und Meth.

irt Phys. Res. A 398 (1997)

.V-64

63

Clearly the spectral brightness of laser-based sources make them superior to synchrotrons for studies of sharp structures in free atoms. As is clear from Table I, the superior peak power available already with. for example, a high-harmonic source as compared to insertion devices makes the use of nonlinear optical processes an attractive and unique possibility. However, a number of technical challenges concerning source optimization, retention of a short pulse duration when the radiation is handled by gratings and other optical components. and optimized focusing optics must be addressed. Many of the techniques now used in the visible and UV spectral region are independent of the wavelength and could in principle be applied in the XUV and soft X-ray region using these new optimized sources. This will enable high-resolution studies of highly charged ions and inner-shell processes. Other spectroscopic investigations not easily carried out with synchrotrons. which become possible with laser-based sources are two-color experiments, allowing the study of ionization from excited states. as well as dynamics of. e.g. fragmentation processes in molecules if the pulse duration matches the timescale of the process. One could think of extending the successful femtochemistry techniques [77], to the UV and VUV domain. thus allowing the dissociation dynamics of many more molecular systems to be studied. Another interesting aspect, specific to high-order harmonics is the possibility of generating trains of attosecond pulses [78.79], recognizing the conceptual similarity between the energetically equidistant high-order harmonics and the cavity modes in a mode-locked CW laser. In addition, by modulating the polarization of the laser light, and using the high sensitivity of harmonic generation to the laser ellipticity. a single attosecond pulse could be selected from such a train of pulses. This area of research is attracting today a lot of attention [SO]. These attosecond pulses have a duration of about 100 as = 10 s. which is of the same order as typical timescales for electron motions in atoms. This opens the route towards ultrafast time-resolved atomic physics.

S. Svanberg. Atomic and Molecular Spectroscopy. 2nd Ed.. Springer Series Atoms and Plasmas. Vol. 6 (Springer. Heidelberg. 1992). Physica Scripta T 34 (1991 ) 77. on Synchrotron Radiation, Vols. Amsterdam. 1983, 1986, 1987 ).

[31 J.T. Costello, J.-P. Mosnier. E.T. Kennedy, P.K. Carroll and
G. OSullivan.

141 E.E. Koch fed.). Handbook

l-3 (North-Holland,
ISI S. Svanberg,

Lol [71 ISI f91

High resolution laser spectroscopy in the UV/ VUV spectral region. in: Applied Laser Spectroscopy. eds. M. lnguscio and W. Demtriider (Plenum Press. New York. 1990) p. 149. J. Larsson and S. Svanberg, Appl. Phys. B 59 ( 1994) 433. B. Wiik. these Proceedings, pp. I - 17. I. Schneider, these Proceedings, pp. 41-53. R. Prazeres. J.-M. Ortega, C. Bazin. M. Bergher. M. Billardon, M.-E. Couprie, M. Velghe and Y. Petroff. Nucl. Instr. and Meth. A 277 (198X) 68.

S. Werin. M. Eriksson. J. Larsson. A. Persson and S. Svanberg. Nucl. Instr. and Meth. A 304 (1991 ) 81. R. Zerne. J. Larsson and S. Svanberg. Phys. Rev. A 49 (1994) 138. D. Strickland and G. Mourou. Opt. Commun. 56 ( 1985) 219. S. Svanberg. J. Larsson. A. Persson and C.-G. Wnhlstriim. Physica Scripta 49 (1994) 1x7. A. LHuillier. T. Auguste, Ph. Balcou. B. Cure. P. Monot. P. Saltere>, C. Altucci. M. Gaarde. J. Larsson, E. Mevel. T. Starczewski. S. Svanberg. C.-G. Wahlstrdm, R. Zeme, KS. Budil. T. Ditmtre and M.D. Perry. Int. J. Nonlinear Opt. 4 ( I995 ) 647. [I51 C.-G. Wahlstrom, in: X-ray Lasers 1994. eds. D.C. Eder and D.L. Matthews (AIP, Woodbury, 1994) p. 312. [I61 C.-G. Wahlstrom. J. Larsson. A. Persson. T. Starczewski. S. Svanberg. P. Salieres, P. Balcou and A. LHuillier. Phys. Rev. A 48 ( 1993 ) 4709. J.J. Mncklin, J.M. Kmetec and C.L. Gordon 111. Phys. Rev.

Lett. 70 ( 1993) 766. M.D. Perry and G. Mourou. Science 264 ( 1994) 917:
A. LHuillier 774. and Ph. Balcou. Phys. Rev Lett. 70 (1993)

1191 S.G. Preston, A. Sanpera, M. Zepf. W.J. Blyth. C.G. Smith,

J.S. Wark. M.H. Key. K. Burnett, M. Nakai. D. Neely and A.A. Offenberger, Phys. Rev. A 53 (1996) 31. [201 J.L. Krause, K.J. Schafer and K.C. Kulander. Phys. Rev. Lett. 68 ( 1992) 3515. [~I1 Y. Akiyama, K. Midorikawa. Y. Matsunawa. Y. Nagata. M. Obara. H. Tashiro and K. Toyoda, Phys. Rev. Lett. 69 (1992) 7176. C.-G. Wahlstrom, S. Borgstrom. J. Larsson and S.-G. Pettersson, Phys. Rev. A Sl ( 1995) 585. C. Altucci, T. Starczewski, C.-G. Wahlstriim, E. Mevel. B. Carre and A. LHuillier. J. Opt. Sot. Am. B I3 ( 1996) 148. J.M. Schins. P. Breger. I? Agoatini. R.C. Constantinescu. H.G. Muller, A. Bouhal. G. Grillon. A. Antonetti and A. Mysyrowtcz. .I. Opt. Sot. Am. B 13 (1996) 197. [251 T. Starczewski, J. Larsson. C.-G. Wahlstriim. M.H.R. Hutchinson, J.E. Muffett. R.A. Smith. and J.W.G. Tisch, J. Phys. B 27 (1994) 3191. f261 T.E. Clover. R.W. Schoenlein, A.H. Chin and C.V. Shank, Phys. Rev. Lett. 76 ( 1996) 1468. T. Ditmire. K.S. Budil. M.D. Perry and A. ~271 P. Sal&s, LHuillier. J. Phys. B 26 (1994) L217.

Acknowledgements The authors acknowledge a most stimulation collaboration with past and present colleagues and graduate students. This work was supported by the Swedish Natural Science Research Council and the Knut and Alice Wallenberg Foundation.

References

[I] W. Demtroder,
Heidelberg.

Laser 1996).

Spectroscopy,

2nd

Ed.

(Springer,

64 1281 P. Sal&es, A. LHuillier Lett. 75 ( 1995) 3776. and M. Lewenstein, Phyh. Rev. [S3] Se. Johansson. A. JoueiLadeh. LT. LltzPn. J. Lars\on, A. Persson, C.-G. WahlstrOm. S. Svanberg. D.S. Leckronr and G. Wahlgren, Astron. J. 421 ( 1994) 809. 1-1 J. Larsaon, R. Zerne and H. Lundberg. J. Phys. B 79 ( 1996) I. [S5] C. Luo. U. Berzinsh. R. Zrrne and S. Svanberg. Phq\. Rr\. A 52 (1995) 1936. [561 U. Berzinsh. C. Luo. R. Zernr, S. Svanberg and E. Biemont, Phyn. Rev. A 55 (March 1997). 1571 J. Larsson. R. Zerne. A. Persson. C.-G. Wahlatriim and S. Svanberg. Z. Phys. D 17 ( 1993) 329. R. [.sSl Zerne. C. Luo. Z. Jiang. J. Larsson and S. Svanberg. Z. Phys. D 37 (1994) 187. [591 M.B. Gaarde. R. Zerne. C. Luo. Z. Jiang, J. Larsson and S. Svanberg, Phys. Rev. 50 ( 1994) 309. E. Mevel. R. Zerne, A. LHuillier, C.-G. [fa J. Larson. Wahlstr&n and S. Svanbrrg, J. Phyr. B 18 ( 1995) L52. I611 S. Svanberg. P. Tsekerih and W. Happer. Phys. Rev. Lett. 30 (197.1) 817. Phys. Rev. LQl S. Haroche. J.A. Palsner and A.L. Schawlow, Lett. 30 ( 1973) 94X. J. Larsson and S. Svanberg. Phyh. Rev. A 31 th31 J. Ben&son. ( 1990) 5457. [641 J. Bengtsson. J. Larsson, S. Svanberg and C.-G. Wahlstriim. Phys. Rev. A 41 (1990) 233. G.J. Bengtsson. P. JGnsson. J. Larsxon and S. Svanberg. Z. (6.51 Phys. D 71 (1991) 437. R. lf561 Teets. J. Eckstein and T.W. Hansch. Phys. Rev. Lett. 38 ( 1977) 760. Phys. Rev. Lett. 38 1671 M.M. Salour and C. Cohen-Tannoudji. (1977) 757. [@I R.G. Beauaoleil. D.H. McIntyre. C.J. Foot, E.A. Hildum, B. Couillaud and T.W. Hansch. Phys. Rev. A 35 ( 1987) 3878: ibid. A 39 (1989) 4591. R. Kallenhack and T.W. HHnxh. Phys. 1691 C. Zimmermann. Rev. Lett. 65 ( 1990) 571. 1701 J.C. Bergquist. F. Diedrich. W.M. Itsno and D.J. Wineland, in: Laser Spectroscopy IX. eds. M.S. Feld. J.E. Thomas and A. Mooradian (Academic Press. Orlando, 1989) p. 274: H. Hemmeti. J.C. Bergquist and W.M. Itano. Opt. Lett. 8 (1983) 73. [711 K.S.E. Eikema, W. Ubachs. W. Waasen and W. Hogervorst, Phys. Rev. Lett. 71 (1993) 1690. K.S.E. Eikema, W. Ubachs and W. Waxen. 1721 W. Hogervorst. in: Laser Spectroscoy. eds. M. Inguscio. M. Allegrini and A. Sass0 (World Scientific. Singapore. 1996) p. 92. [731 C. Wieman and T.W. Hansch. Phys. Rev. A 22 ( 1980) 191. [741 Ph. Balcou. P. Saliires. K.S. Budil. T. Ditmire. M.D. Perry and A. LHmilier. Z. Phys. D 34 (1995) 107. [751 R. Haight and D.R. Peale. Phys. Rev. Lrtt. 70 ( 1993) 3979: R. Haight, Surf. Sci. Rep. 31 (199.5) 275; R. Haight and P.F. Seidler. Appl. Phys. Lett. 65 ( 1994) 5 17. Rep. Prog. Phys. ( 19911 [761 8. Sonntag and P. Zimmermann. 91 I. Vols. I and 2 (World Sci[771 A.H. Zewail, Femtochemistry. entific, Singapore. 1994). [781 P.B. Corkum. N.H. Burnett and M.Y. Ivanov, Opt. Lett. 19 i 1994) 1870. [791 M.Y. Ivanov. P.B. Corkum. T. Zuo and A. Bandrauk. Phys. Rev. Lett. 74 ( 199.5) 2933. I. Mercer. E. Mevel. R. Zerne. A. LHuillier, Ph. Antoine VW and C.-G. Wahlstriim. Phyr. Rev. Lett. 77 ( 1996) 1731.

[?91 R. Danelius. A. Piskarskas, A. Persson Lith. J. Phys. 33 (1993) 328.

[301 H.

and S. Svanherg.

Eichmann. S. Meyer, K. Riepl. C. Momma and B. Wellegehausen. Phys. Rev. A SO ( 1994) R2834. [3]1 M.B. Gaarde. P. Antoine, A. Persson, B. Ca&. A. LHuillier and C.-G. Wahlstriim. J. Phys. B 39 t 1996) Ll63. [3?] J. Larsson. to appear. 1331 R. Zerne. Time-resolved studies of atoms and ions in the <ho&wavelength region, Ph.D. Thesis. LRAP-195 (Lund Institute of Technology, Lund, 1996). 1331 D.L. Matthews, P.L. Hagelstein, M.D. Rosen. M.J. Eckart. N.M. Ceglio, A.U. Hari, H. Medecki, B.J. MacGowan, J.E. Trehes. B.L. Whitten. E.M. Campbell. C.W. Hatcher, A.M. Hawryluk, R.L. Kauffman, L.D. Pleasance. G. Ramhach. J.H. Scotield. G. Stone and T.A. Weaver. Phys. Rev. Lett. 54 (1985) 110. L.B. Da Silva, D.J. Fields, C.J. Keane, J.A. Koch. R.A. London, D.L. Matthews, S. Maxon. S. Mrowka. A.L. Osterheld. J.H. Scofield, G. Smimkaveg. J.E. Trehes and R.S. Walling, Phys. Flulds B 4 (1991) 3316. [361 B.E. Lemoff. G.Y. Yin, C.L. Gordon III. C.P.J. Barty and S.E. Harris, Phys. Rev. Lett. 75 ( 1995) 1574. M. Obara. H. Tashiro and K. [371 Y. Nagata. K. Midorikawa. Toyoda. Phys. Rev. Lett. 71 ( 1993) 3774. 1381 S. BorgstrGm, E. Fill, J. Larsson. T. Starczewshi, J. Steingruber. S. Svanberg and C.-G. Wahlstrbm, Inst. Phys. Conf. Ser. No. I40 (IOP Publishing. Bristol, 1995) p. 141.
[391 B.N. [351 B.J. MacGowan,

Chlchkov. A. Egbert. Nolte and B. Wellegehausen,

H. Eichmann, C. Momma. S. Phys. Rev. A 52 (1995) 1639.

F.G. Tomasel, O.D. Cortazar. D. Hartshorn and J.L.A. Chilla, Phys. Rev. Lett. 73 I 1994) 1191. 1411 E.E. Fill (ed.). X-Ray Lasers 1992 (IOP Publishing, Bristol. 1992). [421 D.E. Eder and D.L. Matthews (eda.), X-Ray Lasers 1994 tAIP Publishing. Woodbury, 1994). L431 S. Svanherg and C.-G. Wahlstriim teds.). X-Ray Lasers 1996 (IOP Publishing, Bristol, 1996). [441 R.A. Lacy, A.C. Nilsson. R.L. Byer. W.T. Silvfast. O.R. Wood II and S. Svanberg. J. Opt. Sot. Am. B 6 ( 1989) 1309. [451 J.D. Kmetec. C.L. Gordon III. J.J. Macklin, B.E. Lemoff. G.S. Brown and S.E. Harrlc, Phys. Rev. Lett. 68 (1993) 1.527. [461 K. Herrlin. G. Svahn, C. Olsson. H. Petterson. C. Tillman. A. Persaon, C.-G. WahlstrGm and S. Svanherg. Radiology I89 c1993) 65. 1471 C. Tillman. A. Persson, C.-G. Wahlstro;m. S. Svanberg and K. Herrlin. Appl. Phys. B 61 (1995) 333. [481 C. Tillman. 1. Mercer. S. Svanberg and K. Herrlin, J. Opt. Sot. Am. B I3 (1996) 1. 1491 C.L. Gordon 111. G.Y. Yin, B.E. Lemoff. P.B. Bell and C.P.J. Batty, Opt. Lett. 20 (1995) 1056. J.M. Schins. P. Berger, P. Agostini, R.C. Constantinescu, L.501 H.G. Muller, G. Grillon. A. Antonetti and A. Mysyrowic7. Phys. Rev. Lett. 73 (1994) 2180. LSll F. Raksi, K.R. Wilson. Z. Jiang, A. Ikhlef, C.Y. CGttt and J.C. Kieffer. J. Chem. Phys. 104 ( 1996) 15. [=I G.J. Bengtsson. U. Berzinsh, J. Larsson and S. Svanberg, Astron. Astrophys. 263 ( 1992) 440.

r401 J.J. Rocca, V.N. Shlyapstev,