- Email: [email protected]
Highly charged ion-induced potential electron emission from clean Au(1 1 1): Dependence on the projectile angle of incidence
NIM B Beam Interactions with Materials & Atoms
Nuclear Instruments and Methods in Physics Research B 256 (2007) 520–523 www.elsevier.com/locate/nimb
Highly charged ion-induced potential electron emission from clean Au(1 1 1): Dependence on the projectile angle of incidence W. Meissl
a,*
a
, M.C. Simon a, J.R. Crespo Lo´pez-Urrutia b, H. Tawara b, J. Ullrich b, HP. Winter a, F. Aumayr a
Institut fu¨r Allgemeine Physik, Technische Universita¨t Wien, A-1040 Wien, Austria b Max-Planck Institut fu¨r Kernphysik, D-69029 Heidelberg, Germany Available online 30 December 2006
Abstract Total yields for electron emission from an atomically clean Au(1 1 1) surface bombarded by slow highly charged Arq+ (q 6 17) and Xe ions (q 6 50) have been measured as a function of ion impact angle (h = 10–70) with respect to the surface normal. For fixed ion impact velocity the electron yields which are dominated by potential electron emission increase with the inverse square root of cos(h). We demonstrate that for given ion species the impact velocity component normal to the surface is the key parameter for potential electron emission from metal targets. 2006 Elsevier B.V. All rights reserved. q+
PACS: 34.50.Dy; 79.20.Rf Keywords: Electron emission; Highly charged ions; Electron beam ion trap; Clean surfaces
1. Introduction Impact of slow ions (impact velocity < 1 a.u. = 25 keV/ amu) on solid surfaces is of genuine interest in plasma- and surface physics and related applications. Nature and intensity of the resulting inelastic processes depend both on the kinetic and the potential (= internal) ion energy carried towards the surface. For most practical applications as, e.g. ion-induced kinetic electron emission (KE [1–4]) or ion–surface scattering and kinetic sputtering [5,6], the kinetic projectile energy is of primary importance. However, ion-induced processes may also depend on the internal (potential) projectile energy, in particular if the latter exceeds the kinetic projectile energy. This will give rise to further electron emission and for certain target materials also to a special kind of sputtering (potential electron emission PE [4,7–10], potential sputtering [11–13]). *
Corresponding author. Tel.: +43 1 58801 13437; fax: +43 1 58801 13499. E-mail address: [email protected] (W. Meissl). 0168-583X/$ - see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2006.12.102
Considerable potential energy is being stored in a highly charged ion (HCI) Zq+ during its production when q electrons are removed from an originally neutral particle. This potential energy is released again when the HCI is neutralized at a solid surface. According to the currently accepted scenario [9,13], HCI neutralization starts with the formation of a transient, multiply-excited particle with some empty inner shells, which has been called ‘‘hollow atom’’ (HA). This name was introduced by Briand et al. [14] in order to illustrate pictorially an electronic structure leading to the appearance of projectile-characteristic soft X-ray emission upon HCI surface impact. Such X-rays arise primarily in the late stage of the HA decay already inside the target bulk, whereas the largest part of the slow electrons will be emitted by autoionization (AI) still before the HA has touched the surface [9]. For metal target surfaces, the above surface part of this interaction can be well described by the so-called ‘‘classical over-barrier model’’ (COB) [15]. As soon as the HA approaches closely the surface, it will be screened by the metal electron gas which further accelerates its deexcitation. Inside the solid the until now surviving
W. Meissl et al. / Nucl. Instr. and Meth. in Phys. Res. B 256 (2007) 520–523
inner shell vacancies of the strongly screened HA will recombine by emission of projectile-characteristic fast Auger electrons and/or soft X-rays, depending on the respective fluorescence yield. The just mentioned projectile recombination and relaxation processes cannot be simply distinguished from each other, since the fast Auger electron emission may already start before close surface contact, and slow electron emission will also continue after penetration of the surface. However, the slow electrons provide information on the HA development above and at the target surface, and fast Auger electron and/or soft X-ray spectra on the HA history mainly below the surface [9,16]. For ions with kinetic energies well above the respective KE threshold [1], the total electron yield will result both from PE and KE, but the relative importance of the two contributions is difficult to assess. An unambiguous measurement of the PE yield is however possible under certain well-defined circumstances. (a) For ions with kinetic energies well below the KE threshold (exclusive PE) [17]; (b) or slow ions in such high charge states that the potential energy greatly exceeds the kinetic energy (dominant PE) [18]; (c) for grazing incidence of fast HCI under the assumption that only the ion velocity component normal to the target surface (‘‘vertical impact velocity’’) is relevant for PE [19]. So far, the validity of assumption (c) could only be demonstrated for ions with relatively low charge states up to Ar8+ [20]. In this contribution we discuss the angular dependence of the PE yield for a clean Au(1 1 1) surface bombarded by HCI in much higher charge states, up to q = 50. It is shown for the first time that also for such highly charged ions assumption (c) remains valid.
521
mental chamber via a 90 analyzing magnet for separation of ions with desired mass-to-charge numbers, after passing an electrostatic quadrupole triplet lens, a switching magnet and a focusing and steering unit. The beamline has a pressure of 2 · 1010 mbar. The ion beam then enters a differentially pumped UHV chamber (base pressure 3 · 1010 mbar) via a 1 mm diam. aperture, followed by a somewhat wider electron repeller aperture. The ions then hit a clean single-crystalline Au(1 1 1) target (see Fig. 1), which is sputter-cleaned by Ar+ ions from a separate commercial sputter gun at regular intervals. The target crystal is mounted on a X-, Y-, Z- and h manipulator (h is the ion impact angle with respect to the target surface normal which can be varied in 1 steps). To detect the number of emitted electrons per ion impact we have applied a slightly modified version of the electron number statistics (ES) method described by Lemell et al. [22]. Electrons emitted from the ion–surface impact region are extracted by a weak electric field through a highly transparent grid and accelerated onto a surface barrier type detector (Canberra PD 100-12-300 AM) biased at +25 kV. 3D ray-trace calculations performed for this geometry with the program SIMION showed that the negative bias of the electron repeller aperture actually assists in collecting the emitted electrons, in particular for small impact angles h. For this reason the negative bias is set to a relatively high value of 450 V. Assuming that the large majority of emitted electrons during HCI impact on a metal surface have energies below 50 eV with an angular distribution following a cosine law [1], SIMION was used to calculate the electron collection efficiency for every given ion impact angle geometry. The collection efficiency at 10 ion impact angle was determined to 85 ± 2%. With increasing impact angle the collection efficiency is rising and reaches 100% at 45. These values were verified experimentally and a corresponding correction was made to all our data. Since the ion beam has a diameter of 1–2 mm
2. Experimental method HCI utilized in the present study have been obtained from an electron beam ion trap (EBIT) at the Max Planck Institut fu¨r Kernphysik in Heidelberg. Details of this EBIT can be found elsewhere [21]. In its continuous mode of operation, also referred to as leaky or dc mode, ions with sufficient thermal energy can leave the trap over an electrostatic barrier provided by a confining drift tube held at constant potential with respect to the central drift tube. The bias of the confining drift tube defines the kinetic energy of the extracted ions (in our case highly charged Arq+ and Xeq+ ions) and has been adjusted for every measurement to keep ions with different charge q at the same velocity of 7 · 105 m/s (2.5 keV/amu). We have set the electron beam current to a value of typically 330 mA for the present measurements. The extracted ions are continuously replenished by injection of neutral gas atoms. The ion beam is focused by an Einzel lens and transported to the experi-
Fig. 1. Experimental setup for measuring ion-induced electron emission statistics. Electron trajectories calculated by SIMION for electrons emitted from the surface with an energy of 20 eV (angular range from 90 to +90) are shown as an illustrative example.
522
W. Meissl et al. / Nucl. Instr. and Meth. in Phys. Res. B 256 (2007) 520–523
depending on its collimation, and the target is a circular single crystal with a diameter of 8 mm, impact angles h larger than 75 result in the ion beam hitting the edge of the target and possibly also the target holder. This drastically changes the electron yield. Data for impact angles below 10 have also not been obtained because then the collection efficiency of our ES setup was unsatisfactory. Electron emission induced by single projectiles finishes within a time interval of well below 1 ps after the collision, much shorter than the time resolution of the applied detector electronics. Thus, n electrons emitted during a particular ion impact and accelerated towards the surface barrier detector (biased at +25 kV) are registered like one electron of n Æ 25 keV rather than n individual 25 keV electrons. The number of electrons emitted for a particular ion impact event can therefore be evaluated from the detector pulse height distributions. More details on the ES detection method and the appropriate correction of measured pulse height spectra can be found in [22,23] and references therein. 3. Results and discussion Total electron yields c for Arq+(q = 11,17) and Xeq+ (26 6 q 6 50) ions impinging with constant velocity v of 7 · 105 m/s (2.5 keV/amu) on a clean single crystal Au(1 1 1) surface are plotted in Fig. 2 versus the ion impact angle h with respect to the target surface normal. First of all, we note that the electron yield strongly increases with the charge state q of the incident ions, as clearly expected for PE [24,25]. Secondly, for a given charge state q the yield increases with the projectile impact angle h. Our experimental data for Arq+ and Xeq+ projectiles follow quite well the inverse square root of cos(h): c2 cðhÞv¼const ¼ c1 þ pffiffiffiffiffiffiffiffiffiffiffiffiffi ; ð1Þ cosðhÞ
Fig. 2. Dependence of total electron yields c (data points) versus impact angle h with respect to the surface normal for impact of highly charged Ar and Xe ions on a clean Au(1 1 1) surface. Impact velocity was fixed at 7 · 105 m/s for all charge states. The solid lines represent fits according to Eq. (1). Typical error bars are shown for q = 48.
with c1 and c2 as free fit parameters. In earlier investigations for normal incident (h = 0) HCI on a polycrystalline Au surface, an empirical relationship for the velocity dependence of electron yields exclusively caused by PE has been found [17,24,25]: c cðvÞh¼0 ¼ pffiffiffi þ c1 ; ð2Þ v c and c1 are free fit parameters. Relationship (2) is well supported by COB modeling calculations [24,26] and reflects the fact that the interaction time available for autoionization and slow electron emitting processes decreases with increasing projectile speed. The interaction path length is the distance between the point where the first electrons are captured and the target surface. We mention that within this picture the velocity independent part of the PE yield c1 corresponds to emission processes taking place at and shortly after the projectile ion has hit the surface. Then electrons still residing in highly excited Rydberg states of the hollow atom will be peeled off [15,24,27], and empty inner shells will be rapidly filled in the first monolayers of the surface, leading to the emission of a correspondingly large number of electrons. The number of slow electrons can well exceed the original ion charge state q, because of repeated HA decay and regeneration [24–26]. It is of interest to test whether the observed impact-angle dependence of the electron emission yields as shown in Fig. 2 is simply determined by the projectile velocity component normal to the surface: v? ¼ v cosðhÞ:
ð3Þ
We have therefore performed impact-angle dependent measurements as shown in Fig. 2 with Xe44+ projectile ions for six different impact energies varied from 2.2 keV/amu to 4.2 keV/amu. All measured Xe44+ data follow surprisingly well the relation c cðv? Þ ¼ pffiffiffiffiffi þ c1 ; ð4Þ v? which simply follows from Eqs. (1)–(3). They can be fitted with one single curve using a c1 fit value of 79.5 e/ion. Results from all Xe44+ measurements are plotted together in Fig. 3(a) and (b) as a function of v?, the projectile velocity component normal to the surface. Before we draw a final conclusion, however, we need to check whether KE could play any role in our experiments, since these have certainly been performed well above the respective KE thresholds of 6 · 104 m/s [28]. Unfortunately, no literature values for the KE yields for impact of neutral or singly charged Ar and Xe projectiles on Au are available for an impact velocity of 7 · 105 m/s. Eder et al. [28,29] have measured KE yields for Ar+ and Xe+ at impact velocities of up to 4 · 105 m/s and 1.5 · 105 m/s, respectively, on polycrystalline Au. Assuming a linear increase with projectile velocity as seen from data for Ar+ and Xe+ on polycrystalline aluminium [30], we extrapolated these KE yields up to 8 · 105 m/s. Based on this
W. Meissl et al. / Nucl. Instr. and Meth. in Phys. Res. B 256 (2007) 520–523
523
Transnational Access granted by the European Project RII3 026015. References
Fig. 3. (a) Measured total electron yields c (data points) versus projectile velocity component normal to the surface v? for impact of Xe44+ on a clean Au(1 1 1) surface, measured at six different impact velocities v (from 2.2 keV/amu to 4.2 keV/amu). The solid line represents a fit according to Eq. (4). (b) Double logarithmic plot of (a) where a constant part c1 = 79.5 e/ion was subtracted from all yields. The straight line with a slope of 1/2 demonstrates the validity of Eq. (4).
approximation, the relevant KE yields in our experiments should remain well below 3 e/ion for both Ar and Xe ions, which is up to two orders of magnitude below the measured total electron yields. Thus, the reported yields are exclusively representative for PE. We therefore conclude that the here observed angular dependence of the PE yields directly reflects the variation of the velocity component normal to the target surface v? (relation (3)) as a function of the impact angle. This clearly demonstrates that processes responsible for emission of slow (Ee < 50 eV) electrons are already completed in the first monolayers of the Au surface, since only under these circumstances the PE yield clearly depends on the interaction time elapsed and thus on v?. So far, this had only been shown for charge states up to q = 8 [20], but now could be demonstrated also for much higher ion charge states up to q = 50. Acknowledgements This work has been supported by Austrian Science Foundation FWF (Project No. 17449) and was carried ¨ AW. The experiout within Association EURATOM-O ments were performed at the distributed ITS-LEIF-Infrastructure at MPI Heidelberg Germany, supported by
[1] D. Hasselkamp, in: G. Ho¨hler (Ed.), Particle Induced Electron Emission II, Vol. 123, Springer, Heidelberg, 1992, p. 1. [2] J. Schou, Scanning Microscopy 2 (1988) 607. [3] M. Ro¨sler, W. Brauer, in: G. Ho¨hler (Ed.), Particle Induced Electron Emission I, Vol. 122, Springer, Berlin, 1991. [4] R. Baragiola, in: J.W. Rabalais (Ed.), Chap. IV in Low Energy Ion– Surface Interactions, Wiley, 1993. [5] P. Sigmund (Ed.), Book Fundamental Processes in Sputtering of Atoms and Molecules (SPUT 92), Mat. Fys. Medd., Copenhagen, 1993. [6] H. Gnaser, Low-Energy Ion Irradiation of Solid Surfaces, Springer, Berlin, 1999. [7] H.D. Hagstrum, Phys. Rev. 96 (1954) 325. [8] H.D. Hagstrum, Phys. Rev. 96 (1954) 336. [9] A. Arnau, F. Aumayr, P.M. Echenique, M. Grether, W. Heiland, J. Limburg, R. Morgenstern, P. Roncin, S. Schippers, R. Schuch, N. Stolterfoht, P. Varga, T.J.M. Zouros, HP. Winter, Surf. Sci. Rep. 27 (1997) 113. [10] H. Winter, Phys. Rep. 367 (2002) 387. [11] T. Neidhart, F. Pichler, F. Aumayr, HP. Winter, M. Schmid, P. Varga, Phys. Rev. Lett. 74 (1995) 5280. [12] M. Sporn, G. Libiseller, T. Neidhart, M. Schmid, F. Aumayr, HP. Winter, P. Varga, M. Grether, N. Stolterfoht, Phys. Rev. Lett. 79 (1997) 945. [13] F. Aumayr, HP. Winter, Phil. Trans. Roy. Soc. (London) 362 (2004) 77. [14] J.P. Briand, L. de Billy, P. Charles, S. Essabaa, P. Briand, R. Geller, J.P. Desclaux, S. Bliman, C. Ristori, Phys. Rev. Lett. 65 (1990) 159. [15] J. Burgdo¨rfer, P. Lerner, F.W. Meyer, Phys. Rev. A 44 (1991) 5674. [16] HP. Winter, F. Aumayr, J. Phys. B: At. Mol. Opt. Phys. 32 (1999) R39. [17] F. Aumayr, H. Kurz, D. Schneider, M.A. Briere, J.W. McDonald, C.E. Cunningham, HP. Winter, Phys. Rev. Lett. 71 (1993) 1943. [18] D. Schneider, M.W. Clark, B. Penetrante, J. McDonald, D. DeWitt, J.N. Bardsley, Phys. Rev. A 44 (1991) 3119. [19] I.G. Hughes, J. Burgdo¨rfer, L. Folkerts, C.C. Havener, S.H. Overbury, M.T. Robinson, D.M. Zehner, P.A. Zeijlmans van Emmichhoven, F.W. Meyer, Phys. Rev. Lett. 71 (1993) 291. [20] C. Lemell, J. Sto¨ckl, J. Burgdo¨rfer, G. Betz, HP. Winter, F. Aumayr, Phys. Rev. Lett. 81 (1998) 1965. [21] J.R. Crespo Lo´pez-Urrutia, B. Bapat, B. Feuerstein, D. Fischer, H. Lo¨rch, R. Moshammer, J. Ullrich, Hyperfine Interact. 146/147 (2003) 109. [22] C. Lemell, J. Sto¨ckl, HP. Winter, F. Aumayr, Rev. Sci. Instr. 70 (1999) 1653. [23] F. Aumayr, G. Lakits, HP. Winter, Appl. Surf. Sci. 47 (1991) 139. [24] H. Kurz, F. Aumayr, C. Lemell, K. To¨glhofer, HP. Winter, Phys. Rev. A 48 (1993) 2182. [25] H. Kurz, F. Aumayr, D. Schneider, M.A. Briere, J.W. McDonald, HP. Winter, Phys. Rev. A 49 (1994) 4693. [26] H. Kurz, F. Aumayr, C. Lemell, K. To¨glhofer, HP. Winter, Phys. Rev. A 48 (1993) 2192. [27] C. Lemell, HP. Winter, F. Aumayr, J. Burgdo¨rfer, F.W. Meyer, Phys. Rev. A 53 (1996) 880. [28] H. Eder, F. Aumayr, HP. Winter, Nucl. Instr. and Meth. B 154 (1999) 185. [29] H. Eder, M. Vana, F. Aumayr, HP. Winter, Rev. Sci. Instr. 68 (1997) 165. [30] B. Svensson, G. Holme´n, J. Appl. Phys. 52 (1981) 6928.
Nuclear Instruments and Methods in Physics Research B 256 (2007) 520–523 www.elsevier.com/locate/nimb
Highly charged ion-induced potential electron emission from clean Au(1 1 1): Dependence on the projectile angle of incidence W. Meissl
a,*
a
, M.C. Simon a, J.R. Crespo Lo´pez-Urrutia b, H. Tawara b, J. Ullrich b, HP. Winter a, F. Aumayr a
Institut fu¨r Allgemeine Physik, Technische Universita¨t Wien, A-1040 Wien, Austria b Max-Planck Institut fu¨r Kernphysik, D-69029 Heidelberg, Germany Available online 30 December 2006
Abstract Total yields for electron emission from an atomically clean Au(1 1 1) surface bombarded by slow highly charged Arq+ (q 6 17) and Xe ions (q 6 50) have been measured as a function of ion impact angle (h = 10–70) with respect to the surface normal. For fixed ion impact velocity the electron yields which are dominated by potential electron emission increase with the inverse square root of cos(h). We demonstrate that for given ion species the impact velocity component normal to the surface is the key parameter for potential electron emission from metal targets. 2006 Elsevier B.V. All rights reserved. q+
PACS: 34.50.Dy; 79.20.Rf Keywords: Electron emission; Highly charged ions; Electron beam ion trap; Clean surfaces
1. Introduction Impact of slow ions (impact velocity < 1 a.u. = 25 keV/ amu) on solid surfaces is of genuine interest in plasma- and surface physics and related applications. Nature and intensity of the resulting inelastic processes depend both on the kinetic and the potential (= internal) ion energy carried towards the surface. For most practical applications as, e.g. ion-induced kinetic electron emission (KE [1–4]) or ion–surface scattering and kinetic sputtering [5,6], the kinetic projectile energy is of primary importance. However, ion-induced processes may also depend on the internal (potential) projectile energy, in particular if the latter exceeds the kinetic projectile energy. This will give rise to further electron emission and for certain target materials also to a special kind of sputtering (potential electron emission PE [4,7–10], potential sputtering [11–13]). *
Corresponding author. Tel.: +43 1 58801 13437; fax: +43 1 58801 13499. E-mail address: [email protected] (W. Meissl). 0168-583X/$ - see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2006.12.102
Considerable potential energy is being stored in a highly charged ion (HCI) Zq+ during its production when q electrons are removed from an originally neutral particle. This potential energy is released again when the HCI is neutralized at a solid surface. According to the currently accepted scenario [9,13], HCI neutralization starts with the formation of a transient, multiply-excited particle with some empty inner shells, which has been called ‘‘hollow atom’’ (HA). This name was introduced by Briand et al. [14] in order to illustrate pictorially an electronic structure leading to the appearance of projectile-characteristic soft X-ray emission upon HCI surface impact. Such X-rays arise primarily in the late stage of the HA decay already inside the target bulk, whereas the largest part of the slow electrons will be emitted by autoionization (AI) still before the HA has touched the surface [9]. For metal target surfaces, the above surface part of this interaction can be well described by the so-called ‘‘classical over-barrier model’’ (COB) [15]. As soon as the HA approaches closely the surface, it will be screened by the metal electron gas which further accelerates its deexcitation. Inside the solid the until now surviving
W. Meissl et al. / Nucl. Instr. and Meth. in Phys. Res. B 256 (2007) 520–523
inner shell vacancies of the strongly screened HA will recombine by emission of projectile-characteristic fast Auger electrons and/or soft X-rays, depending on the respective fluorescence yield. The just mentioned projectile recombination and relaxation processes cannot be simply distinguished from each other, since the fast Auger electron emission may already start before close surface contact, and slow electron emission will also continue after penetration of the surface. However, the slow electrons provide information on the HA development above and at the target surface, and fast Auger electron and/or soft X-ray spectra on the HA history mainly below the surface [9,16]. For ions with kinetic energies well above the respective KE threshold [1], the total electron yield will result both from PE and KE, but the relative importance of the two contributions is difficult to assess. An unambiguous measurement of the PE yield is however possible under certain well-defined circumstances. (a) For ions with kinetic energies well below the KE threshold (exclusive PE) [17]; (b) or slow ions in such high charge states that the potential energy greatly exceeds the kinetic energy (dominant PE) [18]; (c) for grazing incidence of fast HCI under the assumption that only the ion velocity component normal to the target surface (‘‘vertical impact velocity’’) is relevant for PE [19]. So far, the validity of assumption (c) could only be demonstrated for ions with relatively low charge states up to Ar8+ [20]. In this contribution we discuss the angular dependence of the PE yield for a clean Au(1 1 1) surface bombarded by HCI in much higher charge states, up to q = 50. It is shown for the first time that also for such highly charged ions assumption (c) remains valid.
521
mental chamber via a 90 analyzing magnet for separation of ions with desired mass-to-charge numbers, after passing an electrostatic quadrupole triplet lens, a switching magnet and a focusing and steering unit. The beamline has a pressure of 2 · 1010 mbar. The ion beam then enters a differentially pumped UHV chamber (base pressure 3 · 1010 mbar) via a 1 mm diam. aperture, followed by a somewhat wider electron repeller aperture. The ions then hit a clean single-crystalline Au(1 1 1) target (see Fig. 1), which is sputter-cleaned by Ar+ ions from a separate commercial sputter gun at regular intervals. The target crystal is mounted on a X-, Y-, Z- and h manipulator (h is the ion impact angle with respect to the target surface normal which can be varied in 1 steps). To detect the number of emitted electrons per ion impact we have applied a slightly modified version of the electron number statistics (ES) method described by Lemell et al. [22]. Electrons emitted from the ion–surface impact region are extracted by a weak electric field through a highly transparent grid and accelerated onto a surface barrier type detector (Canberra PD 100-12-300 AM) biased at +25 kV. 3D ray-trace calculations performed for this geometry with the program SIMION showed that the negative bias of the electron repeller aperture actually assists in collecting the emitted electrons, in particular for small impact angles h. For this reason the negative bias is set to a relatively high value of 450 V. Assuming that the large majority of emitted electrons during HCI impact on a metal surface have energies below 50 eV with an angular distribution following a cosine law [1], SIMION was used to calculate the electron collection efficiency for every given ion impact angle geometry. The collection efficiency at 10 ion impact angle was determined to 85 ± 2%. With increasing impact angle the collection efficiency is rising and reaches 100% at 45. These values were verified experimentally and a corresponding correction was made to all our data. Since the ion beam has a diameter of 1–2 mm
2. Experimental method HCI utilized in the present study have been obtained from an electron beam ion trap (EBIT) at the Max Planck Institut fu¨r Kernphysik in Heidelberg. Details of this EBIT can be found elsewhere [21]. In its continuous mode of operation, also referred to as leaky or dc mode, ions with sufficient thermal energy can leave the trap over an electrostatic barrier provided by a confining drift tube held at constant potential with respect to the central drift tube. The bias of the confining drift tube defines the kinetic energy of the extracted ions (in our case highly charged Arq+ and Xeq+ ions) and has been adjusted for every measurement to keep ions with different charge q at the same velocity of 7 · 105 m/s (2.5 keV/amu). We have set the electron beam current to a value of typically 330 mA for the present measurements. The extracted ions are continuously replenished by injection of neutral gas atoms. The ion beam is focused by an Einzel lens and transported to the experi-
Fig. 1. Experimental setup for measuring ion-induced electron emission statistics. Electron trajectories calculated by SIMION for electrons emitted from the surface with an energy of 20 eV (angular range from 90 to +90) are shown as an illustrative example.
522
W. Meissl et al. / Nucl. Instr. and Meth. in Phys. Res. B 256 (2007) 520–523
depending on its collimation, and the target is a circular single crystal with a diameter of 8 mm, impact angles h larger than 75 result in the ion beam hitting the edge of the target and possibly also the target holder. This drastically changes the electron yield. Data for impact angles below 10 have also not been obtained because then the collection efficiency of our ES setup was unsatisfactory. Electron emission induced by single projectiles finishes within a time interval of well below 1 ps after the collision, much shorter than the time resolution of the applied detector electronics. Thus, n electrons emitted during a particular ion impact and accelerated towards the surface barrier detector (biased at +25 kV) are registered like one electron of n Æ 25 keV rather than n individual 25 keV electrons. The number of electrons emitted for a particular ion impact event can therefore be evaluated from the detector pulse height distributions. More details on the ES detection method and the appropriate correction of measured pulse height spectra can be found in [22,23] and references therein. 3. Results and discussion Total electron yields c for Arq+(q = 11,17) and Xeq+ (26 6 q 6 50) ions impinging with constant velocity v of 7 · 105 m/s (2.5 keV/amu) on a clean single crystal Au(1 1 1) surface are plotted in Fig. 2 versus the ion impact angle h with respect to the target surface normal. First of all, we note that the electron yield strongly increases with the charge state q of the incident ions, as clearly expected for PE [24,25]. Secondly, for a given charge state q the yield increases with the projectile impact angle h. Our experimental data for Arq+ and Xeq+ projectiles follow quite well the inverse square root of cos(h): c2 cðhÞv¼const ¼ c1 þ pffiffiffiffiffiffiffiffiffiffiffiffiffi ; ð1Þ cosðhÞ
Fig. 2. Dependence of total electron yields c (data points) versus impact angle h with respect to the surface normal for impact of highly charged Ar and Xe ions on a clean Au(1 1 1) surface. Impact velocity was fixed at 7 · 105 m/s for all charge states. The solid lines represent fits according to Eq. (1). Typical error bars are shown for q = 48.
with c1 and c2 as free fit parameters. In earlier investigations for normal incident (h = 0) HCI on a polycrystalline Au surface, an empirical relationship for the velocity dependence of electron yields exclusively caused by PE has been found [17,24,25]: c cðvÞh¼0 ¼ pffiffiffi þ c1 ; ð2Þ v c and c1 are free fit parameters. Relationship (2) is well supported by COB modeling calculations [24,26] and reflects the fact that the interaction time available for autoionization and slow electron emitting processes decreases with increasing projectile speed. The interaction path length is the distance between the point where the first electrons are captured and the target surface. We mention that within this picture the velocity independent part of the PE yield c1 corresponds to emission processes taking place at and shortly after the projectile ion has hit the surface. Then electrons still residing in highly excited Rydberg states of the hollow atom will be peeled off [15,24,27], and empty inner shells will be rapidly filled in the first monolayers of the surface, leading to the emission of a correspondingly large number of electrons. The number of slow electrons can well exceed the original ion charge state q, because of repeated HA decay and regeneration [24–26]. It is of interest to test whether the observed impact-angle dependence of the electron emission yields as shown in Fig. 2 is simply determined by the projectile velocity component normal to the surface: v? ¼ v cosðhÞ:
ð3Þ
We have therefore performed impact-angle dependent measurements as shown in Fig. 2 with Xe44+ projectile ions for six different impact energies varied from 2.2 keV/amu to 4.2 keV/amu. All measured Xe44+ data follow surprisingly well the relation c cðv? Þ ¼ pffiffiffiffiffi þ c1 ; ð4Þ v? which simply follows from Eqs. (1)–(3). They can be fitted with one single curve using a c1 fit value of 79.5 e/ion. Results from all Xe44+ measurements are plotted together in Fig. 3(a) and (b) as a function of v?, the projectile velocity component normal to the surface. Before we draw a final conclusion, however, we need to check whether KE could play any role in our experiments, since these have certainly been performed well above the respective KE thresholds of 6 · 104 m/s [28]. Unfortunately, no literature values for the KE yields for impact of neutral or singly charged Ar and Xe projectiles on Au are available for an impact velocity of 7 · 105 m/s. Eder et al. [28,29] have measured KE yields for Ar+ and Xe+ at impact velocities of up to 4 · 105 m/s and 1.5 · 105 m/s, respectively, on polycrystalline Au. Assuming a linear increase with projectile velocity as seen from data for Ar+ and Xe+ on polycrystalline aluminium [30], we extrapolated these KE yields up to 8 · 105 m/s. Based on this
W. Meissl et al. / Nucl. Instr. and Meth. in Phys. Res. B 256 (2007) 520–523
523
Transnational Access granted by the European Project RII3 026015. References
Fig. 3. (a) Measured total electron yields c (data points) versus projectile velocity component normal to the surface v? for impact of Xe44+ on a clean Au(1 1 1) surface, measured at six different impact velocities v (from 2.2 keV/amu to 4.2 keV/amu). The solid line represents a fit according to Eq. (4). (b) Double logarithmic plot of (a) where a constant part c1 = 79.5 e/ion was subtracted from all yields. The straight line with a slope of 1/2 demonstrates the validity of Eq. (4).
approximation, the relevant KE yields in our experiments should remain well below 3 e/ion for both Ar and Xe ions, which is up to two orders of magnitude below the measured total electron yields. Thus, the reported yields are exclusively representative for PE. We therefore conclude that the here observed angular dependence of the PE yields directly reflects the variation of the velocity component normal to the target surface v? (relation (3)) as a function of the impact angle. This clearly demonstrates that processes responsible for emission of slow (Ee < 50 eV) electrons are already completed in the first monolayers of the Au surface, since only under these circumstances the PE yield clearly depends on the interaction time elapsed and thus on v?. So far, this had only been shown for charge states up to q = 8 [20], but now could be demonstrated also for much higher ion charge states up to q = 50. Acknowledgements This work has been supported by Austrian Science Foundation FWF (Project No. 17449) and was carried ¨ AW. The experiout within Association EURATOM-O ments were performed at the distributed ITS-LEIF-Infrastructure at MPI Heidelberg Germany, supported by
[1] D. Hasselkamp, in: G. Ho¨hler (Ed.), Particle Induced Electron Emission II, Vol. 123, Springer, Heidelberg, 1992, p. 1. [2] J. Schou, Scanning Microscopy 2 (1988) 607. [3] M. Ro¨sler, W. Brauer, in: G. Ho¨hler (Ed.), Particle Induced Electron Emission I, Vol. 122, Springer, Berlin, 1991. [4] R. Baragiola, in: J.W. Rabalais (Ed.), Chap. IV in Low Energy Ion– Surface Interactions, Wiley, 1993. [5] P. Sigmund (Ed.), Book Fundamental Processes in Sputtering of Atoms and Molecules (SPUT 92), Mat. Fys. Medd., Copenhagen, 1993. [6] H. Gnaser, Low-Energy Ion Irradiation of Solid Surfaces, Springer, Berlin, 1999. [7] H.D. Hagstrum, Phys. Rev. 96 (1954) 325. [8] H.D. Hagstrum, Phys. Rev. 96 (1954) 336. [9] A. Arnau, F. Aumayr, P.M. Echenique, M. Grether, W. Heiland, J. Limburg, R. Morgenstern, P. Roncin, S. Schippers, R. Schuch, N. Stolterfoht, P. Varga, T.J.M. Zouros, HP. Winter, Surf. Sci. Rep. 27 (1997) 113. [10] H. Winter, Phys. Rep. 367 (2002) 387. [11] T. Neidhart, F. Pichler, F. Aumayr, HP. Winter, M. Schmid, P. Varga, Phys. Rev. Lett. 74 (1995) 5280. [12] M. Sporn, G. Libiseller, T. Neidhart, M. Schmid, F. Aumayr, HP. Winter, P. Varga, M. Grether, N. Stolterfoht, Phys. Rev. Lett. 79 (1997) 945. [13] F. Aumayr, HP. Winter, Phil. Trans. Roy. Soc. (London) 362 (2004) 77. [14] J.P. Briand, L. de Billy, P. Charles, S. Essabaa, P. Briand, R. Geller, J.P. Desclaux, S. Bliman, C. Ristori, Phys. Rev. Lett. 65 (1990) 159. [15] J. Burgdo¨rfer, P. Lerner, F.W. Meyer, Phys. Rev. A 44 (1991) 5674. [16] HP. Winter, F. Aumayr, J. Phys. B: At. Mol. Opt. Phys. 32 (1999) R39. [17] F. Aumayr, H. Kurz, D. Schneider, M.A. Briere, J.W. McDonald, C.E. Cunningham, HP. Winter, Phys. Rev. Lett. 71 (1993) 1943. [18] D. Schneider, M.W. Clark, B. Penetrante, J. McDonald, D. DeWitt, J.N. Bardsley, Phys. Rev. A 44 (1991) 3119. [19] I.G. Hughes, J. Burgdo¨rfer, L. Folkerts, C.C. Havener, S.H. Overbury, M.T. Robinson, D.M. Zehner, P.A. Zeijlmans van Emmichhoven, F.W. Meyer, Phys. Rev. Lett. 71 (1993) 291. [20] C. Lemell, J. Sto¨ckl, J. Burgdo¨rfer, G. Betz, HP. Winter, F. Aumayr, Phys. Rev. Lett. 81 (1998) 1965. [21] J.R. Crespo Lo´pez-Urrutia, B. Bapat, B. Feuerstein, D. Fischer, H. Lo¨rch, R. Moshammer, J. Ullrich, Hyperfine Interact. 146/147 (2003) 109. [22] C. Lemell, J. Sto¨ckl, HP. Winter, F. Aumayr, Rev. Sci. Instr. 70 (1999) 1653. [23] F. Aumayr, G. Lakits, HP. Winter, Appl. Surf. Sci. 47 (1991) 139. [24] H. Kurz, F. Aumayr, C. Lemell, K. To¨glhofer, HP. Winter, Phys. Rev. A 48 (1993) 2182. [25] H. Kurz, F. Aumayr, D. Schneider, M.A. Briere, J.W. McDonald, HP. Winter, Phys. Rev. A 49 (1994) 4693. [26] H. Kurz, F. Aumayr, C. Lemell, K. To¨glhofer, HP. Winter, Phys. Rev. A 48 (1993) 2192. [27] C. Lemell, HP. Winter, F. Aumayr, J. Burgdo¨rfer, F.W. Meyer, Phys. Rev. A 53 (1996) 880. [28] H. Eder, F. Aumayr, HP. Winter, Nucl. Instr. and Meth. B 154 (1999) 185. [29] H. Eder, M. Vana, F. Aumayr, HP. Winter, Rev. Sci. Instr. 68 (1997) 165. [30] B. Svensson, G. Holme´n, J. Appl. Phys. 52 (1981) 6928.