Experiments in Nuclear Astrophysics

Experiments in Nuclear Astrophysics

Nuclear Physics A 787 (2007) 289c–298c Experiments in Nuclear Astrophysics K. E. Rehma a Physics Division, Argonne National Laboratory, 9700 South C...

377KB Sizes 1 Downloads 10 Views

Nuclear Physics A 787 (2007) 289c–298c

Experiments in Nuclear Astrophysics K. E. Rehma a

Physics Division, Argonne National Laboratory, 9700 South Cass Av., Argonne, IL 60439, USA Astrophysics is a rapidly growing field, driven by observations from ground and spacebased telescopes, which cover all wavelengths, from radio waves to ultra-high energy gamma rays. Nuclear physics experiments provide critical input parameters, which are important for the interpretation of the observational data. New developments in accelerators and instrumentation during the last years have produced many experimental cross sections, masses and half-lives, needed to describe the new observations. In this contribution recent accomplishments relevant to studies of supernovae will be discussed.

1. INTRODUCTION Astrophysical observations performed during the last decade have influenced physics as never before. The search for dark matter, weakly interacting material, which is required to describe the movement of stars in galaxies, has contributed considerably to the development of new detectors and has pushed sensitivity limits into new territory. The observation of an accelerated expansion of the universe, which started with studies of type Ia supernovae, brought us the concept of dark energy. The need to better understand all these astrophysical phenomena has also led to a renaissance of nuclear astrophysics. In terrestrial laboratories we have started to measure nuclear processes at energies which are typical of the conditions encountered in stellar environments [1]. The development of radioactive beams during the last 15 years [2] has allowed us to measure processes that occur during explosive scenarios, such as novae or supernovae explosions. This synergy between observations and other fields of physics has made nuclear astrophysics an exciting area of research. In this contribution I will discuss recent progress in nuclear astrophysics involving nuclei located on the neutron-rich side of the valley of stability. The results of experiments in quiescent as well as in explosive nuclear burning involving nuclei on the proton-rich side of the mass valley are discussed in the following contributions [3,4]. Supernovae explosions are powered either by the strongest (nuclear) or the weakest (gravitational) force of Nature. They are grouped into two classes [5]: A type Ia supernova originates from the explosion of a so-called white dwarf, whose mass, through 0375-9474/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.nuclphysa.2006.12.045


K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c

accretion from a neighboring companion star, exceeds the Chandrasekar limit (M=1.4M ) and explodes in a gigantic thermonuclear explosion, leaving no remnant behind. Type II supernovae are powered by the gravitational collapse of a massive (M≥8M ) stars leaving behind either a neutron star or a black hole. Both types of supernovae produce an explosion with about the same energy output (∼1053 erg), of which 99% are emitted in neutrinos. The remaining 1% is still sufficient to make these events bright enough that they can compete with the total light output of their host galaxies. The occurrence of a supernova is a quite rare event averaging to about 2-3 per century in a given galaxy. Improvements in detection techniques and automation, and the need to study large samples of type Ia supernovae for dark energy research have increased the average detection rates to about one supernova per day. Type II supernovae are important contributors to the production of heavy elements in Nature and are crucial for understanding the abundance of the elements. Nuclear physics plays a role in both types of supernovae via different reactions which I will discuss briefly below.

2. TYPE Ia SUPERNOVAE Type Ia supernovae explosions, occurring on carbon and oxygen-rich white dwarfs, are powered by the energy release from fusion reactions between carbon and oxygen nuclei [6]. Empirically the light curves of type Ia supernovae are found to be very similar, which has led to their use as ’calibrated’ standard candles for dark energy research. In this context it should be noted, however, that none of the fusion cross sections critical to the energy production in type Ia supernovae, have been measured down to the energy regime relevant for these explosions. This is shown in Fig.1 in a plot of the S-factor (i.e. the energy-weighted fusion cross section, corrected for the Coulomb penetrability) vs. energy for the system 12 C+ 16O [7]. The data points [9–11] extend down to about 4 MeV, while the energy typical of temperatures occurring in these events is about 2.5 MeV, as indicated by the arrow. In order to obtain the cross sections at these low energies one has to rely on extrapolations, which are based on different theoretical models as shown by the various lines in Fig.1. It is interesting to note that all extrapolations originating from the optical model[12] predict S-factors that are rising in the astrophysically important low-energy region where so far no experimental data exist. The thick solid line is the result of a different extrapolation technique, which includes the so-called fusion hindrance effect that has been observed in several heavy ion reactions with medium mass ions and targets (A=28-100) [8]. This hindrance results in a maximum of the S-factor, whose location follows a systematic trend. While for these heavier systems (which all have negative Q-values) fusion hindrance must occur at the lowest energies, because of the phase-space restriction at energies E≤-Q, the C+C and C+O reactions with positive Q-values are not necessarily subject to these restrictions. Many experiments, however, point to the need for fusion hindrance in light systems as well. This remains an open question, which can only be settled by new experimental data, especially for the critical 12C + 12 C reaction. This system has been studied by several groups more than 25 years ago down to Ecm ∼2.5 MeV where the cross sections are typically in the nb range[11,13–16]. The importance of these fusion reactions in nuclear astrophysics clearly

K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c

Figure 1. S-factor for the fusion of See text for details.





O in comparison with several extrapolations.

warrants a set of new experiments. With improvements in detectors as well as in accelerator technology, and a better understanding of other critical experimental issues, such as carbon buildup on the target, as discussed e.g. in Ref. [16], it should be possible to push down the cross section limits by another one or two orders of magnitude.

3. TYPE II SUPERNOVAE While the underlying physics of type Ia supernovae explosions is at least qualitatively understood, there are many more open questions for type II supernovae. Starting with a massive star (M≥8M ), the collapse has to be treated relativistically, in three dimensions and including rotations. In addition, since most of the energy is released in the form of neutrinos, the correct neutrino-nucleus cross sections have to be used in the calculations. All these requirements make the study of type II supernovae a complex, multi-faceted problem. Frome the abundance curves of the elements it was concluded that about 50% of the heavy nuclei are produced in an explosive, neutron-rich environment, and type II supernovae are the prime candidates for the site of the so-called r-process, where starting from certain seed nuclei, through a series of neutron-capture reactions followed by β decays, heavier nuclei up to 208Pb and beyond are being formed. The properties of nuclei along the r-process path strongly effect this process. Examples of some recent experimental accomplishments in measurements of half-lives, masses or neutron capture rates will be discussed below. The influence of weak interaction rates (electron capture and β decays) on the dynamics of supernovae explosions are reviewed in the contribution to this conference by R. Zegers [17]. Recent results of neutrino-induced nucleosynthesis can be found in Ref. [18]. Because all nuclei in the r-process path have to be produced during the time scale of the


K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c

explosion, the sum of their β-decay half-lives has to be commensurate with the explosion time. For this reason the r-process follows a path through short-lived, exotic, neutron-rich nuclei, which for the most part has not been accessible to experiments in the past. In order to pin down the path of the r-process the half-lives of nuclei, especially along the closed neutron shells at N=50, 82 and 126 as well as their mass values are needed. Furthermore a small neutron capture rate for nuclei in the vicinity of the closed-neutron shells can lead to an increased impedance of the reaction flow. For this reason neutron capture rates of neutron-rich nuclei are needed. Since the last Conference on Nucleus-Nucleus Collisions in 2003 [19] considerable progress has been made in all three areas.

3.1. Half-life Measurements of Neutron-rich Nuclei Depending on the temperature and the neutron density during the explosion, the rprocess can follow different routes. It is thought to proceed through the closed shell nuclei 78Ni and 82Ge along the N=50 shell before reaching N=82 at 124Mo. It leaves the N=82 shell in the vicinity of 130Cd and reaches the neutron number N=126 around 188Sm. While most of these nuclei, especially the heaviest ones are outside the range of present experiments, considerable progress has been made for nuclei along the N=50 and 82 shell closures. For the N=82 shell a series of experiments at ISOLDE [20] has pushed the half-life measurements down to 129Ag. Life time measurements along the N=50 line are now complete through a recent experiment determining the half-life of the doubly-magic nucleus 78Ni [21]. Using a primary beam of 140 MeV/u 86Kr eleven events 78Ni nuclei were detected resulting in a half-life of 110+100 −60 ms. Together with improvements in the calculations of β decay rates this new experimental value led to a shortening of the duration of the r-process, increasing the yields of the heavy nuclei that can be produced. While for the N=82 closed neutron shell 129Ag is still the lightest nucleus with an experimentally determined half-life, considerable progress has been made for half-life measurements of neighboring neutron-rich nuclei [22].

3.2. Mass Measurements of Neutron-rich Nuclei As shown in Ref. [21] half-life measurements of exotic nuclei can be performed with only 10 events. A mass measurement is the next complex experiment, requiring typically about ten times more nuclei. For this reason the number of mass measurements of rprocess nuclei is very small and with present techniques is limited to the N=50 region. A recent review of these studies can be found in Ref. [23], which also includes a description of the main techniques used in these experiments. While mass measurements of most of the nuclei located directly in the r-process path are presently outside our experimental capabilities, extensions of mass measurements to nuclei located between the valley of stability and the r-process path can already provide valuable information. Recent experiments on neutron-rich Zr and Mo [24] and Ba and La [25] isotopes have shown that neutron-rich nuclei such as 110Mo are less bound than predicted by the 2003 Mass Compilation [26]. If this trend continues, the r-process path might be closer to the valley of stability than assumed in the present calculations.

K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c


3.3. Measurements of (n,γ) Cross Sections The short half-lives of the r-process nuclei make it impossible to use them as targets in (n,γ) experiments. Thus, these neutron capture experiments have to be performed in inverse kinematics with the short-lived nuclei as a beam. However, since neutron targets do not exist, surrogate reactions have to be used, with the (d,p) reactions being the prime candidates. These reactions, performed by bombarding deuterium targets (e.g. CD2 ) with heavy ion beams and detecting the outgoing protons in highly pixilated Si detector arrays, provide many challenges, especially with regard to the Q-value resolution that can be achieved. Many new experiments of inverse (d,p) reactions with radioactive beams have been performed during the last few years [27–37]. Among the reactions studied so far, the 82Ge(d,p)83Ge reaction was the first to involve a nucleus located in the r-process path [32]. The experiment, performed with a 4.6 s 82 Ge beam obtained through protoninduced fission of 238U, exemplifies the difficulties encountered in such measurements. Due to the choice of the (ISOL) production method the 82 Ge beam has contaminants from neighboring isobars (in this case 82 As and 82 Se), whose contribution can be eliminated by measuring the outgoing protons in coincidence with the heavy reaction products. The kinematics of inverse reactions results in low energies of the outgoing protons and a compression of the excitation energy spectrum, leading to a Q-value resolution of only ∼300 keV. This is insufficient to resolve the first excited state in 83 Ge, which is found to be at an excitation energy of Ex = 280±20 keV. The contributions to the energy resolution come from various sources, including the energy loss in the target, the beam spot size, the detector geometry and the need to use inverse kinematics. In this area a new, highresolution, large-acceptance magnetic spectrometer would allow many experiments which presently can not be performed [38]. In addition the question about the relation between (n,γ) capture and (d,p) reactions requires further studies. Some experiments have been performed [39,40] with stable beams in the mass 130 region, but a full understanding has not yet been achieved.

4. INDIVIDUAL REACTION RATES 4.1. The Production of 44 Ti While the majority of nuclei produced in a supernova explosion are radioactive, only a few have sufficiently long half-lives and a suitable gamma decay energy to be detected by orbiting gamma-ray satellites. Examples of these nuclei include 26 Al, 44Ti, 56Ni, 56,57Co and 60 Fe. A comparison between the isotopic amounts deduced from satellite measurements and the amounts calculated in different network calculations can give valuable information about the explosion mechanism. A prerequisite is that the nuclear reaction rates for the production and the destruction of this particular isotope are experimentally well determined. Taking the data from the light curve of SN 1987A [41] or the strength of the 1.157 MeV line observed by the COMPTEL telescope for the SN remnant Cassiopeia A [42] one obtains about twice as much 44 Ti as calculated with the accepted reaction rates. Although many reactions play a role in the production of 44 Ti (T1/2 =59 y) in Type II


K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c

SNe, the major production reaction is 40 Ca(α, γ)44 Ti. Cross sections for this reaction are available in the literature from prompt γ-ray spectroscopy studies using a 40Ca target and a high intensity α beam [43]. This reaction has recently been re-visited with two different techniques. In an experiment performed at the ATLAS accelerator at Argonne [44], 44Ti was produced by bombarding a 4He gas target with a beam of 40Ca. The 44Ti ions were implanted into a Cu catcher, which was later analyzed for its 44Ti content using the technique of Accelerator Mass Spectrometry (AMS). In a second experiment the (inverse) reaction 4 He(40Ca,44Ti)γ was studied by detecting the 44Ti ions in the focal plane of the recoil mass spectrometer DRAGON [45]. While the AMS experiment gives an energy-integrated yield, the recoil separator experiment provides a detailed excitation function. The resulting resonance strength from Ref. [44] is found to be 2-3 times as large as the one obtained by prompt γ-ray spectroscopy. Using this new reaction rate, both the absolute 44Ti yield as well as the 44Ti/56 Ni ratio are now much closer to the values inferred from the γ-ray astronomy data.

4.2. The 12C(α, γ)16 O Reaction This reaction, which has sometimes been called the ’holy grail’ of nuclear astrophysics, plays an important role in the production of carbon and oxygen, two of the most important elements for the creation of life in the Cosmos. The C/O ratio is also a critical parameter for the development of massive stars, determining their fate for ending either as a neutron star or a black hole. Because of the small cross section of the 12C(α, γ)16O reaction at typical red-giant conditions (∼10−17 b) a direct measurement of this reaction is beyond present capabilities. One has therefore to rely on extrapolations, starting from measurements above ∼890 keV, or use indirect techniques. Due to its importance in nuclear astrophysics many studies of the 12C(α, γ)16O reaction have been performed in the past. Recent experiments include measurements of the 16O recoil nuclei with recoil separators [46,47], measurements of the γ-rays with Ge and BaF2 detectors [48,49], as well as indirect techniques [50]. There are two main capture modes for the 12C(α, γ)16O reaction. One is capture with multipolarity E1 with contributions from the 1− resonance at Ex =9.585 MeV and a subthreshold 1− state at Ex =7.117 MeV. The other is E2 capture including contributions from direct processes and the tail from a sub-threshold 2+ state at Ex =6.917 MeV. As shown in Ref. [55], measurements of the beta-delayed alpha decay of 16N provide the tightest constraint on the E1 component of the S-factor S(E1). In these experiments S(E1) is extracted from the height of a small satellite peak located at Eα ∼0.8 MeV, which originates from the interference of the two 1− states mentioned above. In earlier experiments using Si detectors the large β/α ratio (∼ 105 ) created a considerable background at low energies, which partially masked the low-energy part of the α spectrum from which S(E1) is obtained. We have recently re-measured the beta-delayed alpha decay of 16N using a set of highacceptance twin-ionization chambers [51] with minimum thickness, which are practically insensitive to β particles. These detectors measure, in addition to the energy, also the

K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c


Figure 2. Left: Energy spectrum (upstream (2) vs. downstream (1) detector), measured with one of the twin ionization chambers for a foil with implanted 16N particles. Right: Same, but for a non-implanted foil.

emission angles of the two coincident particles (12C and α), which, through the back-toback requirement, provides an additional experimental condition for background suppression. The ionization chambers were calibrated with 7Li-α and t-α coincidences obtained from the reactions 10B(n,α)7 Li and 6 Li(n, α)t, respectively, which give α particles in the energy range of 1.4-2 MeV. The accuracy and long-term stability of these detectors was found to be better than ±5 keV. The 16 N beam (E=62 MeV) was produced at the ATLAS In-Flight Facility [52] via the d( N,16 N)p reaction by bombarding a cryogenically cooled gas cell filled with deuterium with a 82 MeV 15 N beam. The 16N particles were then slowed down and stopped in a 17 µg/cm2 thick carbon foil. Details of the experimental setup will be published in an upcoming paper [53]. 15

A spectrum of one of the twin-ionization chambers (upstream detector vs. downstream detector) is shown in Fig.2. The left part of the figure gives the results of a foil irradiated with 16 N particles, while the right panel comes from an identical non-irradiated foil, representing the background contributions . The two main groups in the left panel correspond to 12 C and α coincidences detected in the up- and downstream detectors, respectively. These two groups show a tail caused by energy loss straggling of the low-energy 12 C particles in the capture foil. The group around channel 2000 on the y-axis, is caused by 16 N decays where the 12 C particle is stopped in the target frame, so that only a part of the energy is detected. An α-particle, emitted downstream and hitting the target frame with only a fraction of the α energy detected in the twin-ionization chamber, gives rise to the group near to the origin. The assignments of the various groups is well reproduced by Monte Carlo calculations. Events where only part of the energy is detected in the


K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c

Figure 3. Preliminary energy spectrum of the β-delayed α decay of 16N obtained in this experiment in comparison with previous measurements (solid and dashed lines). See text for details.

ion-chamber can be eliminated by using the angle information. Consequently, by using ionization chambers for the detection of the α-12 C coincidences, the low-energy satellite peak in the beta-delayed α-spectrum of 16N can now be followed, practically backgroundfree, down to an energy of ∼400 keV. The resulting α-spectrum is shown in the insert of Fig.3. The circles represent the low-energy part of the spectrum. The solid and dashed lines are from Refs. [54,55], respectively. A comparison of our data with the results from Ref. [54] shows good agreement while the spectrum from Ref. [55] has a somewhat shallower slope. The height of the satellite peak at Ecm ∼0.8 MeV, which is sensitive to the S-factor, shows good agreement among the different experiments. The S-factor at Ecm ∼300 keV, SE1(300), is obtained from a fit to the data from the beta-delayed α decay of 16 N, including the results from the capture reaction 12 C(α, γ)16 O at higher energies, as well as phase shift values, obtained from elastic scattering of 12C(α, α). The number of fitting parameters is large and there are strong correlations among some of the parameters. A preliminary value for SE1(300)=73±16 keVb was obtained, which is in good agreement with earlier measurements.

5. SUMMARY Supernovae explosions represent a complex interplay between many fields of physics. They require input from Atomic, Plasma and Nuclear Physics. They have to be treated relativistically, in three dimensions, including turbulent motion, using a network of nuclear reactions, which to a large extent is only based on theoretical estimates. Despite all these shortcomings, our understanding of supernovae has made a major step forward during the

K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c


last decade, driven by advances in observational astronomy as well as in computational technologies. We understand qualitatively the forces behind these events, which are used as ’standard candles’ in cosmological studies. We have started in terrestrial laboratories to study some of the nuclear reactions occuring in these explosive scenarios. The reactions paths during these explosions, however, are still too far away from the valley of stability to be fully covered with present day capabilities. Next generation facilities and detectors are needed to give us a quantitative understanding of these exciting phenomena.

6. Acknowledgment The 16N experiment was performed in collaboration with X. D. Tang, M. Notani, I. Ahmad, C. Brune, A. Champagne, J. Greene, A. Hecht, D. Henderson, R. V. F. Janssens, C. L. Jiang, L. Jisonna, D. Kahl, F. Moore, R. C. Pardo, M. Paul, N. Patel, G. Savard, J. P. Schiffer, R. E. Segel, S. Sinha, B. Shumard and A. Wuosmaa. This work was supported by the US Department of Energy, Nuclear Physics Division, under contract No. W-31-109-ENG-38 and by the NSF Grant No. PHY-02-16783 (Joint Institute for Nuclear Astrophysics).

REFERENCES 1. 2. 3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.


R. Bonetti et al., Phys. Rev. Lett. 82 (1999) 5205. M. S. Smith and K. E. Rehm, Annu. Rev. Nucl. Part. Sci. 51 (2001) 91. S. Kubono, contribution to this conference. H. Schatz, contribution to this conference. A. V. Fillipenko, Annu. Rev. Astron. Astrophys. 35 (1997) 309. W. Hillebrandt and J. C. Niemeyer, Annu. Rev. Astron. Astrophys. 38 (2000) 191. C. L. Jiang et al., Proceedings of the Conference on Reaction Mechanisms and Nuclear Structure at the Coulomb Barrier, 19-23 March 2006, San Servolo, Italy, AIP 2006, p.63, and subm. to Phys. Rev. C. C. L. Jiang et al., Phys. Rev. C89 (2006) 014613. J. R. Patterson et al., Nucl. Phys. A165 (1971) 545 B. Cujec and C. A. Barnes, Nucl. Phys. A266 (1976) 451 P. R. Christensen, Z. E. Switkowski and R. A. Dayers, Nucl. Phys. A280 (1977) 189. W. Fowler, C. Caughlan and B. Zimmermann, Ann. Rev. Astrophys. 13 (1975) 69 J. R. Patterson et al., ApJ 157 (1969) 367. M. Mazarakis et al., Phys. Rev. C7 (1973) 1280. M. D. High and B. Cujec, Nucl. Phys. A282 (1977) 181. E. F. Aguilera et al., Phys. Rev. C73 (2006) 064601. R. Zegers, contribution to this conference. C. Fr¨ohlich et al., Phys. Rev. Lett. 96 (2006) 142502. Proceedings of the Eighth International Conference on Nucleus Nucleus Collisions Moscow, Russia, 17-21 June 2003, ed. Y. Oganessian and R. Kalpakchieva, Nucl. Phys. A734 (2004) 1. K. L. Kratz et al., Proceedings of the International Conference Fission and Properties


21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

50. 51. 52. 53. 54. 55.

K.E. Rehm / Nuclear Physics A 787 (2007) 289c–298c

of Neutron-Rich Nuclei, Nov. 10-15, 1997, Sanibel Island, USA, World Scientific 1998, p.586 P. T. Hosmer et al., Phys. Rev. Lett. 94 (2005) 112501. F. Montes et al., Phys. Rev. C73 (2006) 035801. D. Lunney, Contribution to the 9th Conference Nuclei in the Cosmos, 25-30 June, 2006, Geneva, Switzerland. U. Hager et al., Phys. Rev. Lett. 96 (2006) 042504. G. Savard et al., Intern. J. Mass Spectr. 251 (2006) 252. A. H. Wapstra, G. Audi and C. Thibault, Nucl. Phys. A729 (2003) 129. K. E. Rehm et al., Phys. Rev. Lett. 80 (1998) 676 and Nucl. Instr. Meth. A449 (2000) 208. A. M. Laird et al., Phys. Rev. C66 (2002) 048801. R. Kozub et al., Nucl. Phys. A758 (2005) 753. B. Guo et al., Nucl. Phys. A761 (2005) 162. A. Wuosmaa et al., Phys. Rev. Lett. 94 (2005) 082502. J. Thomas et al., Phys. Rev. C71 (2005) 021302. L. Gaudefroy et al., J. Phys. G31 (2005) S1623. W. Catford et al., J. Phys. G31 (2005) S1655. A. Wuosmaa et al., Phys. Rev. C72 (2005) 061301. A. Obertelli et al., Phys. Lett. B633 (2006) 33. L Gaudefroy et al., Phys. Rev. Lett. 97 (2006) 092501. A. H. Wuosmaa et al., to be published. I. Tomandl et al., Nucl. Phys. A717 (2003) 149. J. Honzatko et al., Nucl. Phys. A756 (2005) 249. R. Diehl, Nucl. Phys. A758 (2005) 225c and references therein. A. F. Iyudin et al., Astron. Astrophys. 284 (1994) L1. E. L. Coopermann et al., Nucl. Phys. A284 (1977) 163. H. Nassar et al., Phys. Rev. Lett. 96 (2006) 041102. C. Vockenhuber et al., Contribution to the 9th Conference Nuclei in the Cosmos, 25-30 June, 2006, Geneva, Switzerland. D. Sch¨ urmann et al., Nucl. Phys. A758 (2005) 367c. L. Buchmann, Nucl. Phys. A758 (2005) 355c. M. Assuncao et al., Phys . Rev. C73 (2006) 055801. R. Plag, M. Heil, F. K¨appeler and K. Wissak, Proceedings of the 12th Conference on Neutron Capture Gamma-Ray Spectroscopy, September 4-9, 2005, Notre Dame, IN, AIP 2006, p. 181 F. Fleurot et al., Phys. Lett. B615 (2005) 167. C. Budtz-Jorgensen et al., Nucl. Instrum. Meth. A258 (1987) 209. B. Harss et al., Rev. Scient. Instrum. 71 (2000) 380. X. D. Tang et al., to be publ. Z. Zhao et al. Phys. Rev. Lett. 70 (1993) 2066. R. Azuma et al., Phys. Rev. C50 (1994) 1194.