- Email: [email protected]
Highly excited states: New experimental windows in photoexcitation
Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
Highly excited states: New experimental windows in photoexcitation Peter Hammond ∗ School of Physics, University of Western Australia, Perth, WA 6009, Australia Available online 17 March 2005
Abstract In a recent series of experiments, it has been demonstrated that highly excited states, formed through photoexcitation using vacuum-ultraviolet (VUV) synchrotron radiation, have a significant probability of decaying via the fluorescence decay route. This decay route has products of VUV fluorescence photons and, in some cases, long lived metastable atoms. In helium, studies using high resolving powers at third generation synchrotron light sources utilising this decay route to the metastable atoms have enabled the identification of three series of triplet doubly-excited states below the N = 2 ionisation threshold which are excited by single photon impact from the singlet ground state. The pulsed nature of synchrotron light and the detection of fluorescence photons has been utilised to enable fluorescence lifetime measurements and the determination of the (2p3d)1 P lifetime. Fluorescence timing measurements have also allowed the nearly-energy-degenerate ion states 32 S, 32 P and 32 D to be separated through their different lifetimes to allow ion state specific cross sections as a function of energy to be extracted. These groups of measurements have stimulated new theoretical work which explicitly includes both the fluorescence and autoionisation decay routes to reveal for some states a branching ratio in favour of fluorescence. © 2005 Elsevier B.V. All rights reserved. Keywords: Photoexcitation; Synchrotron radiation; Helium; Radiative decay; Metastable atoms; Autoionisation; Fluorescence; Lifetime; Doubly-excited states
1. Introduction Atomic and molecular structure is most commonly thought about in terms of the combination of many single electron orbitals. These orbitals result in radial probability densities which describe the charge distribution in the system to roughly concentrate in certain regions in radius from the nuclear core and relate to the shell structure of systems. For many processes, and particularly in the ultra-violet (UV) and vacuum-ultra-violet (VUV), it is only the outermost electrons that are considered optically active, and in many situations involving excitation, only one electron is excited. States in which two electrons are simultaneously promoted to other orbitals are called “doubly-excited” or “highly-excited”. It is these types of state that are the subject of the present studies. Historically, the earliest mention in the literature for the existence of doubly-excited states seems to be by Kruger [1] in 1930 in an analysis of the spectral lines produced in arc discharges in helium. Many years passed, before the ∗
Tel.: +44 8 6488 2746; fax: +44 8 6488 1014. E-mail address: [email protected].
0368-2048/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.elspec.2005.01.292
doubly-excited states of helium were studied in photoabsorption from the ground state by Madden and Codling [2] in 1963 using synchrotron radiation. These photoexcited doubly-excited states were observed as resonances interpreted by Cooper et al. [3] and identified as being of 1 Po character belonging to the series (sp,2n+) converging on the N = 2 ionisation limit at 65.4 eV. The existence of two further series of energetically narrower states (sp,2n−) and (2pnd), both of 1 Po character was predicted. The line shapes of the resonances observed by Madden and Codling were as predicted by Fano [4] resulting from interference between continuum ionisation and the resonant photoexcitation with subsequent autoionisation of the doubly-excited states. Following these studies, experimental efforts were concentrated into photoabsorption, photoion and photoelectron spectroscopy. Photoion spectroscopy came to dominate the study of doubly-excited states partially because of the extremely good measurement sensitivity that could be achieved, and in helium marvellously detailed experiments were performed (e.g. [5,6]). Photoelectron spectroscopy was used less extensively but did provide the capability to study the decay by autoionisation of doubly-excited states into
14
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
Fig. 2. (a) Neutral particle spectra over the energy region of series of doublyexcited states converging on the He+ (N = 2) threshold. (b) Neutral particle spectra recorded with improved statistical accuracy and compared with the photoelectron spectrum (arising from autoionisation) (adapted from [15]). Fig. 1. Secondary ion signal (RHS) and conventional photoion signal (LHS) i the energy regions of (a) N = 2, (b) N = 3 and (c) N = 4 ionisation thresholds in helium (from [15]).
selected final states of the helium ion and termed “constant ion state” (CIS) spectra (e.g. [7]). Theoretical modelling was in very good agreement with photoabsorption, photoion and photoelectron data (e.g. [8,9]). Fluorescence from doubly excited states has been observed in beam foil spectroscopy [10] and following electron excitation [11]. The new sequence of experiments began in late in 1994, when Hammond’s group in Manchester, during the course of making photoelectron spectroscopy measurements of helium [7] at the Daresbury synchrotron radiation source with a resolving power of ∼1000, noticed a peak like structure at the N = 1 ionisation threshold observed in what was thought to be an ion detection system. It was established that this peak was arising from neutral particles leaving the interaction region of gas and photon beams since photoions were prevented from leaving the interaction region by the presence of a weak electric field. Photoelectrons were also excluded by an appropriately biased mesh. Thus, only neutral particles and photons could enter the detector region where a channel electron multiplier, out of line of sight of the interaction region, could be biased to detect either electrons or ions produced by secondary processes within the detector. The right hand side of Fig. 1 shows the structures recorded near the N = 2–4 ionisation thresholds in helium when secondary ions were detected. The relative yield scale applies to these
spectra. The yield of secondary ions arose from the field ionisation of highly excited Rydberg atoms produced in the fluorescence decay of high n doubly-excited states. The left hand side of Fig. 1 shows photoion signal, recorded with no electric field in the interaction region. Overall, the core advantage of this in-direct detection system was the insensitivity to VUV photons. This allowed experiments to be conducted above the He+ (N = 2) ionisation threshold where the fluorescence of excited ions masks the neutral particle signal in line-of-sight detection experiments. A major concern in these early studies was whether the signal recorded was the result of direct excitation or the effects of secondary processes (discussed in detail in [12,13]). The detailed VUV transition measurements from doubly to singly excited states of helium in a microwave discharge by Baltzer and Karlsson [14] significantly contributed to the interpretation of our new spectra as arising from a direct process. More detailed studies were performed, including pressure dependencies, electric field dependencies and time-offlight spectra which were reported at a sequence of conferences and finally in Physical Review Letters [15] and detailed in the Ph.D. thesis of Odling-Smee [13]. Fig. 2 shows a key spectrum recorded during these studies in which, though the energy resolution was ∼60 meV, peaks in both the metastable atom spectrum and the fluorescence photon spectrum can be assigned to the energetically narrow series (sp,2n−)1 P and which have intensities similar to the peaks arising from the (sp,2n+)1 P series. Very significantly the
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
Fig. 3. Schematic diagram of the radiative decay scheme: illustrated for the decay of the (sp,23) 1 P and (2p3d) 1 P doubly-excited states. Transitions labeled “VUV” produce photons that could have been detected. Singly excited states are metastable for high n (from [15]).
spectrum was also interpreted as showing evidence for significant fluorescence photon yield from the (2p3d)1 P state, the energetically lowest member of the (2pnd)1 P series, lying just below the (sp,24−)1 P state. In addition, a fluorescence decay route scheme was proposed, shown in Fig. 3, which linked the doubly-excited states to the singly-excited states and the cascade photons possible from subsequent transitions. The key outcome from these studies is that photoexcitation of highly-excited states in atomic and molecular targets can be explored via both fluorescence decay and autoionisation. Many of the possible decay routes, and the decay products, are illustrated in Fig. 4 for helium. The first high resolving power measurements detecting fluorescence photons from doubly-excited states in helium
15
were made at the third generation synchrotron source Elettra (Trieste, Italy) by Rubensson and co-workers with an energy resolution of a few meV. Their high quality spectra clearly demonstrated the way in which fluorescence decay favoured the detection of the energetically narrower, and hence longer-lived, states belonging to the (sp,2n−) and (2pnd)1 P Rydberg series [16]. The theoretical analysis of these spectra, in combination with results from a subsequent experiment performed with even higher energy resolution [17] showed that radiative and spin-orbit effects play an important role in the overall characterisation of highly excited states. These results and analysis suggested the energy region where doubly-excited states become closely spaced near ionisation thresholds to be of particular interest for the possible observation of new structure in helium. In neon, Lablanquie et al. [18] observed fluorescence from the (2s2p6 np)1 Po states by an indirect measurement using threshold electron spectroscopy with electron/photon coincidence techniques. LS-forbidden highly excited states were observed by Canton-Rogan et al. [19] in photoelectron spectroscopy studies in argon in confirmation of theoretical predictions [20] and made possible by resolving the CIS spectra of the 3p−1 3/2,1/2 ion states. Metastable atom decay products of fluorescence provided perhaps the most surprising observation arising from the new series of experiments. Penent et al. [21], in a further development of the experimental technique of the OdlingSmee et al. detector layout [13,15], used two microchannel plates positioned near the interaction region, both protected by multiple-mesh biased-grids to prevent charged particle detection. One detector was primarily used for metastable atom detection while the other was used for VUV photon detection. In the spectrum, a portion of which is shown in Fig. 5, additional structures associated with series of doubly-excited states, in excess of the theoretically expected three 1 Po series, were observed. These additional series were identified as (2pnd)3 D and (sp2n−)3 Po , with indications of
Fig. 4. Summary of photoexcitation and decay by fluorescence and autoionisation with their decay products in helium.
16
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
Fig. 5. Triplet and singlet doubly-excited states observed in the yield of metastable atoms (lower spectrum). Photoion yield (upper spectrum) (adapted from [21]).
two further states in the spectra ((2p6d)3 Po and (2p7d)3 Po which were confirmed in subsequent studies [23,24]). Triplet doubly-excited states appear very strongly in the metastable atom spectrum shown in Fig. 5. The strongest triplet series (2pnd)3 D lies to the high energy side of the (sp,2n+)1 Po doubly-excited states which appear strongly in the photoion spectrum (upper spectrum in Fig. 5). More recent experiments by the same research group have confirmed these observations and improved upon the experimental technique to show that the triplet states are most easily observed in the metastable atom detection channel. The new states are observed because the spin–orbit interaction mixes them with the optically allowed 1 Po states and they are detected with high sensitivity since the triplet states in decay by fluorescence cascade into the (1s2s)3 S metastable state with a near 100% branching ratio. Since 2001, the energy region below the N = 2 ionisation threshold has been explored in significant detail including the angular distribution of fluorescence [23,24]. For the metastable atom detection channel, these experiments exploited the pulsed nature of the synchrotron light source in a storage ring filling mode in which the pulsed light spaced by 2 ns is produced for a period of ∼500 ns followed by a period of ∼100 ns in which no light pulses are produced. In this “dark gap”, the time gated metastable signal was recorded to enable photon free spectra to be obtained. It is also important to note that the doubly-excited states appear in metastable atom and VUV photon detection channels as symmetric Voigt profiles as can be seen in Fig. 5. These profiles result from the convolution of the Lorenzian natural lineshape with the usually Gaussian lineshape of the incident photon beam. The signal is observed on a small background signal, since unlike conventional photoion measurements, there is no continuum background signal. This type of spectroscopy is, therefore, of great interest to the development, characterisation and optimisation of high energy resolution synchrotron beamlines. Cascade decay route information has been experimentally explored by Schartner et al. [22] by the spectral analysis of
Fig. 6. (a) Photon yield from the He+ 2p ion state and fit giving a lifetime of 102 ± 15 ps. (b) Photon yield from the (2p3d)1 Po doubly-excited state and fit giving a lifetime of 190 ± 30 ps. The solid lines are instrumental response functions of the system (from [28]).
fluorescence. The (2p3d)1 Po → (1s3d)1 De → (1s2p)1 Po → (1s2 )1 Se cascade transitions were observed in the vacuum ultraviolet and the visible spectral range. It was found that the (2pnd) levels populate predominantly (1snd)1 D levels, while for the (sp,2n+ or −) levels, n1 S levels are also populated. These observations confirm the decay route proposals put forward by [15] in the first measurements of the metastable atom and photon fluorescence product detection channels. Members of the (2pnd)1 Po series were not detected until 1992, using photoion spectroscopy using high resolving power synchrotron radiation [25,26], since the states are energetically very narrow and overlap poorly with the incident photon beam profile producing a small resonance signal which is difficult to distinguish from the large continuum background of ions. In fluorescence, members of (2pnd)1 Po series produce a significant yield in a low background signal environment as can be seen from the (2p7d)1 Po state in Fig. 5. Since a synchrotron storage ring is a pulsed light source of ∼60 ps temporal width, a series of experiments were commenced in 1999 to exploit the clarity of the fluorescence signal from the (2p3d)1 Po state to record the photons from fluorescence in the time domain with the intent of measuring the lifetime of the state. The early experiments suffered from an inter-bunch coupling in the storage ring which caused time averaged light pulses to have a double peaked structure (for full details see [27]). Subsequently, these problems were solved by using special storage ring fills being developed for free electron laser operation at Elettra in which the ring was filled with four electron bunches equi-spaced in the ring. The resulting time-resolved fluorescence spectrum is shown in Fig. 6 from which the lifetime of the (2p3d)1 Po state was determined to be 190 ± 30 ps [28]. The figure also shows the time-resolved fluorescence signal from the transition He+ (2p) to He+ (1s) which has a measured lifetime of 99.717 ± 0.075 ps [29]. A straightforward interpretation of the independent particle
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
model would suggest a similar lifetime for the doubly-excited state as for the ion decay for the “same” 2p inner electron. The presence of the 3d electron modifies the behaviour of the inner electron through the role of the electron–electron interaction. The most recent theoretical calculations of the (2p3d)1 Po lifetime are in good agreement with the measureˇ ment, where Liu et al. [30] predict 207 ps while Zitnik et al. [31] determined lifetimes of 216 and 194 ps in length and velocity form calculations, respectively. More recent measurements [32], exploring the lifetime of states in the series from n = 3–9 indicate a lifetime which lengthens with increasing n and in broad agreement with calculation [31]. ˇ The calculations of Zitnik et al. [31] give extensive predictions for cross sections, the decay rates by autoionisation and fluorescence and detail other parameters for singlet and triplet doubly-excited states lying below the N = 2 ionisation threshold. For the (2p3d)1 Po ,state the decay rates (per ns) are for autoionisation 0.182 and for fluorescence 4.64, thus indicating that decay by fluorescence dominates the decay of the state—a factor contributing to making the state difficult to detect in autoionisation decay route experiments. Time-resolved fluorescence measurements have also been extended to include the decay of excited states of the ion reached by direct ionisation and autoionisation processes as indicated in Fig. 4 [33]. In these experiments, the timeresolved signal arises from the sum of the exponential time profiles from each hydrogenic ion state (3s 2 S, 9.7 ns; 3p 2 P, 0.3 ns; 3d 2 D, 0.97 ns; 2s 2 S, 2 ms and 2p 2 P, 0.1 ns)—states of the same N are near degenerate in energy and unresolvable in transition wavelength. The time-resolved signals are recorded over a scale of approximately tens of ns and as a function of incident photon energy. A multi-parameter least squares fitting procedure (see for example [34,27]) is then applied to each timing spectrum for each photon energy and in which decay exponentials are fixed at the specific lifetimes of the ion states, but are varied in relative amplitude. These fitted relative amplitudes as a function of photon energy then give the partial photoionisation cross-section into each ion state. An example of two of these partial cross-sections is shown in Fig. 7 for the 3p 2 P and 3d 2 D ion states in comparison with the theoretical predictions of S´anchez and Martin [35]. The CIS spectra extracted from the time-resolved signal data show good agreement with the theoretical predictions. Stimulated by the observation of metastable atom decay products, electron impact experiments have been developed to isolate the fluorescence decay products of electron excited doubly-excited states using a time-of-flight method [36,37]. These measurements have produced evidence for highly excited states in H2 which are able to decay via the fluorescence decay route. Studies using the autoionisation decay route in helium have also developed in novel directions. Photoexcitation in very high electric fields, ∼80 kV cm−1 , has resulted in complex doubly-excited state structures [38]. The peak like structure at the N = 2 ionisation threshold has been observed in novel angle resolved photoelectron studies at 90◦ to the in-
17
Fig. 7. Partial photoionisation cross-sections for the 3p 2 P and 3d 2 D ion states. Solid curves: theoretical predictions of S´anchez and Martin [35] with arbitrary vertical offset for clarity (adapted from [33]).
cident photon beam polarisation [39]. Photoexcitation of the dipole forbidden 2p2 1 D2 has been observed by measuring the non-dipolar forward backward asymmetry of photoelectron angular distributions [40]. All of the new approaches to study highly excited states represent new windows through which to look the physics of photoexcitation processes in unprecedented detail. The high resolving power of modern third generation synchrotron sources has been crucial to these developments. Measurements in the time-domain, particularly as free electron laser VUV beamlines are developed, will lead to further insights.
Acknowledgements The skill and insights of colleagues from Manchester, Paris and Tsukuba, all named in co-authored papers below, are acknowledged with enthusiasm. The resource support of EPSRC, Daresbury SRS, Elettra, the ALS, KEK Tsukuba, AMRFP, ARC and UWA has been crucial for the achievement of the studies described here.
References [1] [2] [3] [4] [5] [6] [7] [8] [9]
P.G. Kruger, Phys. Rev. 36 (1930) 855. R.P. Madden, K. Codling, Phys. Rev. Lett. 10 (1963) 516. J.W. Cooper, U. Fano, F. Prats, Phys. Rev. Lett. 10 (1963) 518. U. Fano, Phys. Rev. 124 (1961) 1866. K. Schulz, G. Kaindl, M. Domke, J.D. Bozek, P.A. Heimann, A.S. Schlachter, J.M. Rost, Phys. Rev. Lett. 77 (1996) 3086. R. P¨uttner, B. Gr´emaud, D. Delande, M. Domke, M. Martins, A.S. Schlachter, G. Kaindl, Phys. Rev. Lett. 86 (2001) 3747. E. Sokell, A.A. Wills, J. Comer, P. Hammond, J. Phys. B 29 (1996) L863. J.M. Rost, J. Phys. B 30 (1997) 4663. T. Schneider, C.N. Liu, J.M. Rost, Phys. Rev. A 65 (2002) 042715.
18
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
[10] H.G. Berry, J. Desesquelles, M. Dufay, Phys. Rev. A 6 (1972) 600. [11] W.B. Westerveld, F.B. Kets, H.G.M. Heideman, J. van Eck, J. Phys. B 12 (1979) 2575. [12] E. Sokell, A.A. Wills, P. Hammond, M.A. MacDonald, M.K. OdlingSmee, J. Phys. B 29 (1996) L863. [13] M.K. Odling-Smee, Ph.D. Thesis, University of Manchester, 1998. [14] P. Baltzer, L. Karlsson, Phys. Rev. A 38 (1988) 2322. [15] M.K. Odling-Smee, S. Sokell, P. Hammond, M.A. MacDonald, Phys. Rev. Lett. 84 (2001) 2598. [16] J.E. Rubensson, C. S˚athe, S. Cramm, B. Kessler, S. Stranges, R. Richter, M. Alagia, M. Coreno, Phys. Rev. Lett. 83 (1999) 947. [17] T.W. Gorczyca, J.E. Rubensson, C. S˚athe, M. Str¨om, M. Ag˚aker, D. Ding, S. Stranges, R. Richter, M. Alagia, Phys. Rev. Lett. 85 (2000) 1202. [18] P. Lablanquie, F. Penent, R.I. Hall, J.H.D. Eland, P. Bolognesi, D. Cooper, G.C. King, L. Avaldi, R. Camilloni, S. Stranges, M. Coreno, ˇ K.C. Prince, A. M¨uehleisen, M. Zitnik, Phys. Rev. Lett. 84 (2000) 431. [19] S.E. Canton-Rogan, A.A. Wills, T.W. Gorczyca, M. Wiedenhoeft, O. Nayandin, Chien-Nan Liu, N. Berrah, Phys. Rev. Lett. 85 (2000) 3113. [20] C.N. Liu, A.F. Starace, Phys. Rev. A 59 (1999) R1731. ˇ [21] F. Penent, P. Lablanquie, R.I. Hall, M. Zitnik, K. Bucar, S. Stranges, R. Richter, M. Alagia, P. Hammond, J.G. Lambourne, Phys. Rev. Lett. 86 (2001) 2758. [22] K.-H. Schartner, B. Zimmermann, S. Kammer, S. Mickat, H. Schmoranzer, A. Ehresmann, H. Liebel, R. Follath, G. Reichardt, Phys. Rev. A 64 (2001) 040501. [23] J.G. Lambourne, F. Penent, P. Lablanquie, R.I. Hall, M. Ahmad, M. ˇ Zitnik, K. Bucar, P. Hammond, S. Stranges, R. Richter, M. Alagia, M. Coreno, J. Phys. B: At. Mol. Opt. Phys. 36 (2003) 4339.
[24] J.G. Lambourne, F. Penent, P. Lablanquie, R.I. Hall, M. Ahmad, M. ˇ Zitnik, K. Bucar, P. Hammond, S. Stranges, R. Richter, M. Alagia, M. Coreno, J. Phys. B: At. Mol. Opt. Phys. 36 (2003) 4351. [25] M. Domke, et al., Phys. Rev. Lett. 66 (1991) 1306. [26] M. Domke, et al., Phys. Rev. Lett. 69 (1992) 1171. [27] J.G. Lambourne, Ph.D. Thesis, University of Manchester, 2000. [28] J.G. Lambourne, F. Penent, P. Lablanquie, R.I. Hall, M. Ahmad, ˇ M. Zitnik, K. Bucar, M.K. Odling-Smee, J.R. Harries, P. Hammond, D.K. Waterhouse, S. Stranges, R. Richter, M. Alagia, M. Coreno, M. Ferianis, Phys. Rev. Lett. 90 (2003) 153004. [29] G.W.F. Drake, et al., Phys. Rev. A 46 (1992) 113. [30] C. Liu, et al., Phys. Rev. A 64 (2001) 010501. ˇ [31] M. Zitnik, K. Bucar, M. Stuhec, F. Penent, R.I. Hall, P. Lablanquie, Phys. Rev. A 65 (2002) 032520. [32] J.R. Harries, J.P. Sullivan, Y. Azuma, J. Phys. B: At. Mol. Opt. Phys. 37 (2004) L169. [33] J.R. Harries, J.P. Sullivan, S. Obara, P. Hammond, Y. Azuma, J. Phys. B: At. Mol. Opt. Phys. 36 (2003) L319. [34] P. Hammond, J. Phys. B: At. Mol. Opt. Phys. 29 (1996) L231. [35] I. S´anchez, F. Martin, Phys. Rev. A 48 (1993) 1243. [36] A.J. Murray, P. Hammond, Adv. At. Mol. Phys. 47 (2001) 163–204. [37] J.R. Harries, P. Hammond, A.J. Murray, J. Phys. B: At. Mol. Opt. Phys. 36 (2003) 2579. [38] J.R. Harries, J.P. Sullivan, J.B. Sternberg, S. Obara, T. Suzuki, P. Hammond, J. Bozek, N. Berrah, M. Halka, Y. Azuma, Phys. Rev. Lett. 90 (2003) 133002. [39] A.A. Wills, E. Sokell, T.W. Gorczyca, X. Feng, M. Wiedenhoeft, S.E. Canton, N. Berrah, J. Phys. B: At. Mol. Opt. Phys. 35 (2002) L367. [40] B. Kr¨assig, E.P. Kanter, S.H. Southworth, R. Guillemin, O. Hemmers, D.W. Lindle, R. Wehlitz, N.L.S. Martin, Phys. Rev. Lett. 88 (2002) 203002.
Highly excited states: New experimental windows in photoexcitation Peter Hammond ∗ School of Physics, University of Western Australia, Perth, WA 6009, Australia Available online 17 March 2005
Abstract In a recent series of experiments, it has been demonstrated that highly excited states, formed through photoexcitation using vacuum-ultraviolet (VUV) synchrotron radiation, have a significant probability of decaying via the fluorescence decay route. This decay route has products of VUV fluorescence photons and, in some cases, long lived metastable atoms. In helium, studies using high resolving powers at third generation synchrotron light sources utilising this decay route to the metastable atoms have enabled the identification of three series of triplet doubly-excited states below the N = 2 ionisation threshold which are excited by single photon impact from the singlet ground state. The pulsed nature of synchrotron light and the detection of fluorescence photons has been utilised to enable fluorescence lifetime measurements and the determination of the (2p3d)1 P lifetime. Fluorescence timing measurements have also allowed the nearly-energy-degenerate ion states 32 S, 32 P and 32 D to be separated through their different lifetimes to allow ion state specific cross sections as a function of energy to be extracted. These groups of measurements have stimulated new theoretical work which explicitly includes both the fluorescence and autoionisation decay routes to reveal for some states a branching ratio in favour of fluorescence. © 2005 Elsevier B.V. All rights reserved. Keywords: Photoexcitation; Synchrotron radiation; Helium; Radiative decay; Metastable atoms; Autoionisation; Fluorescence; Lifetime; Doubly-excited states
1. Introduction Atomic and molecular structure is most commonly thought about in terms of the combination of many single electron orbitals. These orbitals result in radial probability densities which describe the charge distribution in the system to roughly concentrate in certain regions in radius from the nuclear core and relate to the shell structure of systems. For many processes, and particularly in the ultra-violet (UV) and vacuum-ultra-violet (VUV), it is only the outermost electrons that are considered optically active, and in many situations involving excitation, only one electron is excited. States in which two electrons are simultaneously promoted to other orbitals are called “doubly-excited” or “highly-excited”. It is these types of state that are the subject of the present studies. Historically, the earliest mention in the literature for the existence of doubly-excited states seems to be by Kruger [1] in 1930 in an analysis of the spectral lines produced in arc discharges in helium. Many years passed, before the ∗
Tel.: +44 8 6488 2746; fax: +44 8 6488 1014. E-mail address: [email protected].
0368-2048/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.elspec.2005.01.292
doubly-excited states of helium were studied in photoabsorption from the ground state by Madden and Codling [2] in 1963 using synchrotron radiation. These photoexcited doubly-excited states were observed as resonances interpreted by Cooper et al. [3] and identified as being of 1 Po character belonging to the series (sp,2n+) converging on the N = 2 ionisation limit at 65.4 eV. The existence of two further series of energetically narrower states (sp,2n−) and (2pnd), both of 1 Po character was predicted. The line shapes of the resonances observed by Madden and Codling were as predicted by Fano [4] resulting from interference between continuum ionisation and the resonant photoexcitation with subsequent autoionisation of the doubly-excited states. Following these studies, experimental efforts were concentrated into photoabsorption, photoion and photoelectron spectroscopy. Photoion spectroscopy came to dominate the study of doubly-excited states partially because of the extremely good measurement sensitivity that could be achieved, and in helium marvellously detailed experiments were performed (e.g. [5,6]). Photoelectron spectroscopy was used less extensively but did provide the capability to study the decay by autoionisation of doubly-excited states into
14
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
Fig. 2. (a) Neutral particle spectra over the energy region of series of doublyexcited states converging on the He+ (N = 2) threshold. (b) Neutral particle spectra recorded with improved statistical accuracy and compared with the photoelectron spectrum (arising from autoionisation) (adapted from [15]). Fig. 1. Secondary ion signal (RHS) and conventional photoion signal (LHS) i the energy regions of (a) N = 2, (b) N = 3 and (c) N = 4 ionisation thresholds in helium (from [15]).
selected final states of the helium ion and termed “constant ion state” (CIS) spectra (e.g. [7]). Theoretical modelling was in very good agreement with photoabsorption, photoion and photoelectron data (e.g. [8,9]). Fluorescence from doubly excited states has been observed in beam foil spectroscopy [10] and following electron excitation [11]. The new sequence of experiments began in late in 1994, when Hammond’s group in Manchester, during the course of making photoelectron spectroscopy measurements of helium [7] at the Daresbury synchrotron radiation source with a resolving power of ∼1000, noticed a peak like structure at the N = 1 ionisation threshold observed in what was thought to be an ion detection system. It was established that this peak was arising from neutral particles leaving the interaction region of gas and photon beams since photoions were prevented from leaving the interaction region by the presence of a weak electric field. Photoelectrons were also excluded by an appropriately biased mesh. Thus, only neutral particles and photons could enter the detector region where a channel electron multiplier, out of line of sight of the interaction region, could be biased to detect either electrons or ions produced by secondary processes within the detector. The right hand side of Fig. 1 shows the structures recorded near the N = 2–4 ionisation thresholds in helium when secondary ions were detected. The relative yield scale applies to these
spectra. The yield of secondary ions arose from the field ionisation of highly excited Rydberg atoms produced in the fluorescence decay of high n doubly-excited states. The left hand side of Fig. 1 shows photoion signal, recorded with no electric field in the interaction region. Overall, the core advantage of this in-direct detection system was the insensitivity to VUV photons. This allowed experiments to be conducted above the He+ (N = 2) ionisation threshold where the fluorescence of excited ions masks the neutral particle signal in line-of-sight detection experiments. A major concern in these early studies was whether the signal recorded was the result of direct excitation or the effects of secondary processes (discussed in detail in [12,13]). The detailed VUV transition measurements from doubly to singly excited states of helium in a microwave discharge by Baltzer and Karlsson [14] significantly contributed to the interpretation of our new spectra as arising from a direct process. More detailed studies were performed, including pressure dependencies, electric field dependencies and time-offlight spectra which were reported at a sequence of conferences and finally in Physical Review Letters [15] and detailed in the Ph.D. thesis of Odling-Smee [13]. Fig. 2 shows a key spectrum recorded during these studies in which, though the energy resolution was ∼60 meV, peaks in both the metastable atom spectrum and the fluorescence photon spectrum can be assigned to the energetically narrow series (sp,2n−)1 P and which have intensities similar to the peaks arising from the (sp,2n+)1 P series. Very significantly the
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
Fig. 3. Schematic diagram of the radiative decay scheme: illustrated for the decay of the (sp,23) 1 P and (2p3d) 1 P doubly-excited states. Transitions labeled “VUV” produce photons that could have been detected. Singly excited states are metastable for high n (from [15]).
spectrum was also interpreted as showing evidence for significant fluorescence photon yield from the (2p3d)1 P state, the energetically lowest member of the (2pnd)1 P series, lying just below the (sp,24−)1 P state. In addition, a fluorescence decay route scheme was proposed, shown in Fig. 3, which linked the doubly-excited states to the singly-excited states and the cascade photons possible from subsequent transitions. The key outcome from these studies is that photoexcitation of highly-excited states in atomic and molecular targets can be explored via both fluorescence decay and autoionisation. Many of the possible decay routes, and the decay products, are illustrated in Fig. 4 for helium. The first high resolving power measurements detecting fluorescence photons from doubly-excited states in helium
15
were made at the third generation synchrotron source Elettra (Trieste, Italy) by Rubensson and co-workers with an energy resolution of a few meV. Their high quality spectra clearly demonstrated the way in which fluorescence decay favoured the detection of the energetically narrower, and hence longer-lived, states belonging to the (sp,2n−) and (2pnd)1 P Rydberg series [16]. The theoretical analysis of these spectra, in combination with results from a subsequent experiment performed with even higher energy resolution [17] showed that radiative and spin-orbit effects play an important role in the overall characterisation of highly excited states. These results and analysis suggested the energy region where doubly-excited states become closely spaced near ionisation thresholds to be of particular interest for the possible observation of new structure in helium. In neon, Lablanquie et al. [18] observed fluorescence from the (2s2p6 np)1 Po states by an indirect measurement using threshold electron spectroscopy with electron/photon coincidence techniques. LS-forbidden highly excited states were observed by Canton-Rogan et al. [19] in photoelectron spectroscopy studies in argon in confirmation of theoretical predictions [20] and made possible by resolving the CIS spectra of the 3p−1 3/2,1/2 ion states. Metastable atom decay products of fluorescence provided perhaps the most surprising observation arising from the new series of experiments. Penent et al. [21], in a further development of the experimental technique of the OdlingSmee et al. detector layout [13,15], used two microchannel plates positioned near the interaction region, both protected by multiple-mesh biased-grids to prevent charged particle detection. One detector was primarily used for metastable atom detection while the other was used for VUV photon detection. In the spectrum, a portion of which is shown in Fig. 5, additional structures associated with series of doubly-excited states, in excess of the theoretically expected three 1 Po series, were observed. These additional series were identified as (2pnd)3 D and (sp2n−)3 Po , with indications of
Fig. 4. Summary of photoexcitation and decay by fluorescence and autoionisation with their decay products in helium.
16
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
Fig. 5. Triplet and singlet doubly-excited states observed in the yield of metastable atoms (lower spectrum). Photoion yield (upper spectrum) (adapted from [21]).
two further states in the spectra ((2p6d)3 Po and (2p7d)3 Po which were confirmed in subsequent studies [23,24]). Triplet doubly-excited states appear very strongly in the metastable atom spectrum shown in Fig. 5. The strongest triplet series (2pnd)3 D lies to the high energy side of the (sp,2n+)1 Po doubly-excited states which appear strongly in the photoion spectrum (upper spectrum in Fig. 5). More recent experiments by the same research group have confirmed these observations and improved upon the experimental technique to show that the triplet states are most easily observed in the metastable atom detection channel. The new states are observed because the spin–orbit interaction mixes them with the optically allowed 1 Po states and they are detected with high sensitivity since the triplet states in decay by fluorescence cascade into the (1s2s)3 S metastable state with a near 100% branching ratio. Since 2001, the energy region below the N = 2 ionisation threshold has been explored in significant detail including the angular distribution of fluorescence [23,24]. For the metastable atom detection channel, these experiments exploited the pulsed nature of the synchrotron light source in a storage ring filling mode in which the pulsed light spaced by 2 ns is produced for a period of ∼500 ns followed by a period of ∼100 ns in which no light pulses are produced. In this “dark gap”, the time gated metastable signal was recorded to enable photon free spectra to be obtained. It is also important to note that the doubly-excited states appear in metastable atom and VUV photon detection channels as symmetric Voigt profiles as can be seen in Fig. 5. These profiles result from the convolution of the Lorenzian natural lineshape with the usually Gaussian lineshape of the incident photon beam. The signal is observed on a small background signal, since unlike conventional photoion measurements, there is no continuum background signal. This type of spectroscopy is, therefore, of great interest to the development, characterisation and optimisation of high energy resolution synchrotron beamlines. Cascade decay route information has been experimentally explored by Schartner et al. [22] by the spectral analysis of
Fig. 6. (a) Photon yield from the He+ 2p ion state and fit giving a lifetime of 102 ± 15 ps. (b) Photon yield from the (2p3d)1 Po doubly-excited state and fit giving a lifetime of 190 ± 30 ps. The solid lines are instrumental response functions of the system (from [28]).
fluorescence. The (2p3d)1 Po → (1s3d)1 De → (1s2p)1 Po → (1s2 )1 Se cascade transitions were observed in the vacuum ultraviolet and the visible spectral range. It was found that the (2pnd) levels populate predominantly (1snd)1 D levels, while for the (sp,2n+ or −) levels, n1 S levels are also populated. These observations confirm the decay route proposals put forward by [15] in the first measurements of the metastable atom and photon fluorescence product detection channels. Members of the (2pnd)1 Po series were not detected until 1992, using photoion spectroscopy using high resolving power synchrotron radiation [25,26], since the states are energetically very narrow and overlap poorly with the incident photon beam profile producing a small resonance signal which is difficult to distinguish from the large continuum background of ions. In fluorescence, members of (2pnd)1 Po series produce a significant yield in a low background signal environment as can be seen from the (2p7d)1 Po state in Fig. 5. Since a synchrotron storage ring is a pulsed light source of ∼60 ps temporal width, a series of experiments were commenced in 1999 to exploit the clarity of the fluorescence signal from the (2p3d)1 Po state to record the photons from fluorescence in the time domain with the intent of measuring the lifetime of the state. The early experiments suffered from an inter-bunch coupling in the storage ring which caused time averaged light pulses to have a double peaked structure (for full details see [27]). Subsequently, these problems were solved by using special storage ring fills being developed for free electron laser operation at Elettra in which the ring was filled with four electron bunches equi-spaced in the ring. The resulting time-resolved fluorescence spectrum is shown in Fig. 6 from which the lifetime of the (2p3d)1 Po state was determined to be 190 ± 30 ps [28]. The figure also shows the time-resolved fluorescence signal from the transition He+ (2p) to He+ (1s) which has a measured lifetime of 99.717 ± 0.075 ps [29]. A straightforward interpretation of the independent particle
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
model would suggest a similar lifetime for the doubly-excited state as for the ion decay for the “same” 2p inner electron. The presence of the 3d electron modifies the behaviour of the inner electron through the role of the electron–electron interaction. The most recent theoretical calculations of the (2p3d)1 Po lifetime are in good agreement with the measureˇ ment, where Liu et al. [30] predict 207 ps while Zitnik et al. [31] determined lifetimes of 216 and 194 ps in length and velocity form calculations, respectively. More recent measurements [32], exploring the lifetime of states in the series from n = 3–9 indicate a lifetime which lengthens with increasing n and in broad agreement with calculation [31]. ˇ The calculations of Zitnik et al. [31] give extensive predictions for cross sections, the decay rates by autoionisation and fluorescence and detail other parameters for singlet and triplet doubly-excited states lying below the N = 2 ionisation threshold. For the (2p3d)1 Po ,state the decay rates (per ns) are for autoionisation 0.182 and for fluorescence 4.64, thus indicating that decay by fluorescence dominates the decay of the state—a factor contributing to making the state difficult to detect in autoionisation decay route experiments. Time-resolved fluorescence measurements have also been extended to include the decay of excited states of the ion reached by direct ionisation and autoionisation processes as indicated in Fig. 4 [33]. In these experiments, the timeresolved signal arises from the sum of the exponential time profiles from each hydrogenic ion state (3s 2 S, 9.7 ns; 3p 2 P, 0.3 ns; 3d 2 D, 0.97 ns; 2s 2 S, 2 ms and 2p 2 P, 0.1 ns)—states of the same N are near degenerate in energy and unresolvable in transition wavelength. The time-resolved signals are recorded over a scale of approximately tens of ns and as a function of incident photon energy. A multi-parameter least squares fitting procedure (see for example [34,27]) is then applied to each timing spectrum for each photon energy and in which decay exponentials are fixed at the specific lifetimes of the ion states, but are varied in relative amplitude. These fitted relative amplitudes as a function of photon energy then give the partial photoionisation cross-section into each ion state. An example of two of these partial cross-sections is shown in Fig. 7 for the 3p 2 P and 3d 2 D ion states in comparison with the theoretical predictions of S´anchez and Martin [35]. The CIS spectra extracted from the time-resolved signal data show good agreement with the theoretical predictions. Stimulated by the observation of metastable atom decay products, electron impact experiments have been developed to isolate the fluorescence decay products of electron excited doubly-excited states using a time-of-flight method [36,37]. These measurements have produced evidence for highly excited states in H2 which are able to decay via the fluorescence decay route. Studies using the autoionisation decay route in helium have also developed in novel directions. Photoexcitation in very high electric fields, ∼80 kV cm−1 , has resulted in complex doubly-excited state structures [38]. The peak like structure at the N = 2 ionisation threshold has been observed in novel angle resolved photoelectron studies at 90◦ to the in-
17
Fig. 7. Partial photoionisation cross-sections for the 3p 2 P and 3d 2 D ion states. Solid curves: theoretical predictions of S´anchez and Martin [35] with arbitrary vertical offset for clarity (adapted from [33]).
cident photon beam polarisation [39]. Photoexcitation of the dipole forbidden 2p2 1 D2 has been observed by measuring the non-dipolar forward backward asymmetry of photoelectron angular distributions [40]. All of the new approaches to study highly excited states represent new windows through which to look the physics of photoexcitation processes in unprecedented detail. The high resolving power of modern third generation synchrotron sources has been crucial to these developments. Measurements in the time-domain, particularly as free electron laser VUV beamlines are developed, will lead to further insights.
Acknowledgements The skill and insights of colleagues from Manchester, Paris and Tsukuba, all named in co-authored papers below, are acknowledged with enthusiasm. The resource support of EPSRC, Daresbury SRS, Elettra, the ALS, KEK Tsukuba, AMRFP, ARC and UWA has been crucial for the achievement of the studies described here.
References [1] [2] [3] [4] [5] [6] [7] [8] [9]
P.G. Kruger, Phys. Rev. 36 (1930) 855. R.P. Madden, K. Codling, Phys. Rev. Lett. 10 (1963) 516. J.W. Cooper, U. Fano, F. Prats, Phys. Rev. Lett. 10 (1963) 518. U. Fano, Phys. Rev. 124 (1961) 1866. K. Schulz, G. Kaindl, M. Domke, J.D. Bozek, P.A. Heimann, A.S. Schlachter, J.M. Rost, Phys. Rev. Lett. 77 (1996) 3086. R. P¨uttner, B. Gr´emaud, D. Delande, M. Domke, M. Martins, A.S. Schlachter, G. Kaindl, Phys. Rev. Lett. 86 (2001) 3747. E. Sokell, A.A. Wills, J. Comer, P. Hammond, J. Phys. B 29 (1996) L863. J.M. Rost, J. Phys. B 30 (1997) 4663. T. Schneider, C.N. Liu, J.M. Rost, Phys. Rev. A 65 (2002) 042715.
18
P. Hammond / Journal of Electron Spectroscopy and Related Phenomena 144–147 (2005) 13–18
[10] H.G. Berry, J. Desesquelles, M. Dufay, Phys. Rev. A 6 (1972) 600. [11] W.B. Westerveld, F.B. Kets, H.G.M. Heideman, J. van Eck, J. Phys. B 12 (1979) 2575. [12] E. Sokell, A.A. Wills, P. Hammond, M.A. MacDonald, M.K. OdlingSmee, J. Phys. B 29 (1996) L863. [13] M.K. Odling-Smee, Ph.D. Thesis, University of Manchester, 1998. [14] P. Baltzer, L. Karlsson, Phys. Rev. A 38 (1988) 2322. [15] M.K. Odling-Smee, S. Sokell, P. Hammond, M.A. MacDonald, Phys. Rev. Lett. 84 (2001) 2598. [16] J.E. Rubensson, C. S˚athe, S. Cramm, B. Kessler, S. Stranges, R. Richter, M. Alagia, M. Coreno, Phys. Rev. Lett. 83 (1999) 947. [17] T.W. Gorczyca, J.E. Rubensson, C. S˚athe, M. Str¨om, M. Ag˚aker, D. Ding, S. Stranges, R. Richter, M. Alagia, Phys. Rev. Lett. 85 (2000) 1202. [18] P. Lablanquie, F. Penent, R.I. Hall, J.H.D. Eland, P. Bolognesi, D. Cooper, G.C. King, L. Avaldi, R. Camilloni, S. Stranges, M. Coreno, ˇ K.C. Prince, A. M¨uehleisen, M. Zitnik, Phys. Rev. Lett. 84 (2000) 431. [19] S.E. Canton-Rogan, A.A. Wills, T.W. Gorczyca, M. Wiedenhoeft, O. Nayandin, Chien-Nan Liu, N. Berrah, Phys. Rev. Lett. 85 (2000) 3113. [20] C.N. Liu, A.F. Starace, Phys. Rev. A 59 (1999) R1731. ˇ [21] F. Penent, P. Lablanquie, R.I. Hall, M. Zitnik, K. Bucar, S. Stranges, R. Richter, M. Alagia, P. Hammond, J.G. Lambourne, Phys. Rev. Lett. 86 (2001) 2758. [22] K.-H. Schartner, B. Zimmermann, S. Kammer, S. Mickat, H. Schmoranzer, A. Ehresmann, H. Liebel, R. Follath, G. Reichardt, Phys. Rev. A 64 (2001) 040501. [23] J.G. Lambourne, F. Penent, P. Lablanquie, R.I. Hall, M. Ahmad, M. ˇ Zitnik, K. Bucar, P. Hammond, S. Stranges, R. Richter, M. Alagia, M. Coreno, J. Phys. B: At. Mol. Opt. Phys. 36 (2003) 4339.
[24] J.G. Lambourne, F. Penent, P. Lablanquie, R.I. Hall, M. Ahmad, M. ˇ Zitnik, K. Bucar, P. Hammond, S. Stranges, R. Richter, M. Alagia, M. Coreno, J. Phys. B: At. Mol. Opt. Phys. 36 (2003) 4351. [25] M. Domke, et al., Phys. Rev. Lett. 66 (1991) 1306. [26] M. Domke, et al., Phys. Rev. Lett. 69 (1992) 1171. [27] J.G. Lambourne, Ph.D. Thesis, University of Manchester, 2000. [28] J.G. Lambourne, F. Penent, P. Lablanquie, R.I. Hall, M. Ahmad, ˇ M. Zitnik, K. Bucar, M.K. Odling-Smee, J.R. Harries, P. Hammond, D.K. Waterhouse, S. Stranges, R. Richter, M. Alagia, M. Coreno, M. Ferianis, Phys. Rev. Lett. 90 (2003) 153004. [29] G.W.F. Drake, et al., Phys. Rev. A 46 (1992) 113. [30] C. Liu, et al., Phys. Rev. A 64 (2001) 010501. ˇ [31] M. Zitnik, K. Bucar, M. Stuhec, F. Penent, R.I. Hall, P. Lablanquie, Phys. Rev. A 65 (2002) 032520. [32] J.R. Harries, J.P. Sullivan, Y. Azuma, J. Phys. B: At. Mol. Opt. Phys. 37 (2004) L169. [33] J.R. Harries, J.P. Sullivan, S. Obara, P. Hammond, Y. Azuma, J. Phys. B: At. Mol. Opt. Phys. 36 (2003) L319. [34] P. Hammond, J. Phys. B: At. Mol. Opt. Phys. 29 (1996) L231. [35] I. S´anchez, F. Martin, Phys. Rev. A 48 (1993) 1243. [36] A.J. Murray, P. Hammond, Adv. At. Mol. Phys. 47 (2001) 163–204. [37] J.R. Harries, P. Hammond, A.J. Murray, J. Phys. B: At. Mol. Opt. Phys. 36 (2003) 2579. [38] J.R. Harries, J.P. Sullivan, J.B. Sternberg, S. Obara, T. Suzuki, P. Hammond, J. Bozek, N. Berrah, M. Halka, Y. Azuma, Phys. Rev. Lett. 90 (2003) 133002. [39] A.A. Wills, E. Sokell, T.W. Gorczyca, X. Feng, M. Wiedenhoeft, S.E. Canton, N. Berrah, J. Phys. B: At. Mol. Opt. Phys. 35 (2002) L367. [40] B. Kr¨assig, E.P. Kanter, S.H. Southworth, R. Guillemin, O. Hemmers, D.W. Lindle, R. Wehlitz, N.L.S. Martin, Phys. Rev. Lett. 88 (2002) 203002.