- Email: [email protected]
Formation of sputtered silver clusters under bombardment with SF5+ ions
Nuclear Instruments and Methods in Physics Research B 197 (2002) 43–48 www.elsevier.com/locate/nimb
Formation of sputtered silver clusters under bombardment with SFþ 5 ions S. Ghalab a, C. Staudt a, S.E. Maksimov b, P. Mazarov a, V.I. Tugushev b, N.Kh. Dzhemilev b, A. Wucher a,* a
b
Institute of Experimental Physics, University of Essen, D-45117 Essen, Germany Cluster Physics Laboratory, Arifov Institute of Electronics, 700125 Tashkent, Uzbekistan Received 30 May 2002
Abstract þ þ The formation of Agn clusters and Agþ n cluster ions under bombardment of a silver surface with SF5 and Xe projectile ions was investigated experimentally. In order to obtain information about the relative abundance of clusters among the flux of sputtered particles independent of their charge state, mass spectra of both secondary ions and sputtered neutral particles were recorded. The neutral species were post-ionized prior to mass analysis by means of photo-ionization using an intense UV laser at a wavelength of 193 nm. It is found that measured Agþ n signals increase þ þ significantly if SFþ 5 projectiles are used instead of rare gas (Ar or Xe ) ions of the same kinetic impact energy. The signals of neutral Ag atoms and Agn clusters, on the other hand, exhibit only a relatively small increase, thus indicating that the enhancement observed for the secondary ions is predominantly caused by an increased ionization probability of sputtered particles under SFþ 5 bombardment rather than by enhanced partial sputtering yields. While the transition from Arþ to Xeþ projectiles leads to a drastic increase of the relative abundance of larger clusters in the spectrum, practically no such effect can be detected for the transition from Arþ to SFþ 5 . This finding shows that the use of polyatomic SF5 projectiles does not lead to a higher efficiency in producing sputtered clusters. Ó 2002 Elsevier Science B.V. All rights reserved.
PACS: 79.20.Rf; 36.40.)c
1. Introduction If a solid is bombarded with keV-ions, particles are released from the surface due to elastic collisions (‘‘sputtering’’). It has been known for a long time that the flux of sputtered particles contains –
*
Corresponding author. Tel.: +49-201-183-4141; fax: +49201-183-93-4141. E-mail address: [email protected] (A. Wucher).
besides atomic species – molecules and clusters which may be composed of up to many hundred atoms [1]. The formation of such clusters in sputtering represents an intriguing phenomenon, since it is not at all clear how these species survive the relatively violent collisional processes that lead to the ejection of atoms from the bombarded surface. In general, the typical abundance pattern of clusters observed in sputtering strongly decreases with increasing cluster size. The quantitative relation between the relative abundance and the number of
0168-583X/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 2 ) 0 1 3 6 1 - 7
44
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48
atoms in a sputtered cluster is generally found to closely follow a power law with a decay exponent d that depends on the sputtering conditions [2–4]. As a general trend, it is found that the relative abundance of large clusters increases (and, hence, d decreases) with increasing total sputtering yield (i.e. the average number of atoms released per primary projectile ion) [4,5]. From a number of mass spectrometric experiments on secondary ions, i.e. those sputtered particles that leave the surface in either a positively or negatively charged ionic state, it has been suggested that the use of polyatomic projectiles may significantly enhance the sputtering yield. One prominent example projectile that has been demonstrated to be particularly effective in secondary ion mass spectrometry (SIMS) of organic surface species is SFþ 5 . More specifically, it has been shown that particularly the signal measured for complex molecular ions may be drastically enhanced under SFþ 5 bombardment [6–8]. These observations suggest that the relative abundance of clusters among the sputtered flux may also be enhanced if the surface is bombarded with SFþ 5 ions as compared to rare gas ion projectiles of the same kinetic energy [9]. The present paper is intended to examine this question by investigating the influence of the projectile on the composition of the flux of particles sputtered from a clean metallic surface. As a first step, we use silver as a target material since it is known that rare gas ion bombardment of a silver surface produces relatively large amounts of Agn clusters [10]. Since it is not a priori clear whether the secondary ion yield enhancements observed in the literature are caused by an increased formation or an increased ionization of the ejected species in the course of the sputtering process, it is necessary to investigate both ionic and neutral clusters leaving the surface.
2. Experimental The experiments reported in this work have been performed in collaboration between two laboratories using two different experimental setups. Part of the experimental data was taken with a secondary ion mass spectrometer located at the
Arifov Institute at Tashkent (setup 1) which has been described in detail elsewhere [11]. In short, a double focusing arrangement consisting of a magnetic sector and an electrostatic prism is used to detect ionic clusters which are sputtered from a polycrystalline silver sample under bombardment with Arþ , Xeþ or SFþ 5 ions impinging under 45° with respect to the surface normal. The primary ions are generated by an electron impact ion source with axial symmetric magnetic field delivering total beam currents of about 0.6 lA for Arþ , Xeþ and SFþ 5 ions. In order to permit a rapid switching between different projectile ions, the source was operated with a gas mixture of Argon, Xenon and SF6 , and the desired projectile was selected by means of a Wien filter. The energy of the primary ions was kept constant at 11 keV. By changing the polarity of the ion optical potentials, both positive and negative cluster ions can in principle be detected, the present experiments, however, are restricted to positive secondary ions. The setup is mounted in an ultrahigh vacuum (UHV) chamber with a base pressure of about 109 mbar. During the experiments, the working pressure rises to about 2 107 mbar due to the operation of the gas ion source. The second part of the experiments were performed with a laser post-ionization reflectron time-of-flight mass spectrometer located at the University of Essen (setup 2, base pressure 109 mbar, working pressure 3 109 mbar) which has also been previously described [12–15]. In this setup, a polycrystalline silver sample is bombarded under 45° incidence with Xeþ or SFþ 5 ions of 10 keV. The ion beam is generated by a commercial cold cathode plasma ion source (Atomica Microfocus) that delivers total beam currents of 330 nA (Xeþ ) and 77 nA (SFþ 5 ) when operated with Xenon or SF6 gas under otherwise comparable conditions. During data acquisition, the primary ion beam was operated in a pulsed mode with a pulse length of 10 ls at a repetition rate of 10 Hz. Mass spectra of neutral particles sputtered from the surface are recorded by post-ionization of the ejected neutral species using an intense UV laser pulse. The ionizing radiation was generated by an excimer laser operated at k ¼ 193 nm with a cross-section of about 2.5 mm2 in the interaction region located
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48
45
about 1 mm above the surface. Secondary ion spectra are measured by simply switching the ionization laser off and leaving the remainder of the experiment unchanged. This way, secondary ions and neutrals are detected under otherwise the same experimental conditions [16]. The procedures employed for data acquisition with this instrument have been described in much detail elsewhere [16]. The recorded time-of-flight spectra were averaged over typically several 1000 instrument cycles, i.e. primary ion pulses and – where applicable – laser shots. Note that in this particular experiment the photo-ionization efficiency is not saturated, and the measured signals of post-ionized neutrals therefore strongly depend on the laser power density. The pulse energy of the laser was therefore carefully monitored and kept constant within a few percent by means of an optical feedback loop.
3. Results and discussion Fig. 1 shows mass spectra of post-ionized sputtered neutral atoms and clusters which have been ejected from the polycrystalline silver surface under bombardment with 10-keV Xeþ and SFþ 5 ions under otherwise identical experimental conditions (setup 2). The different traces depicted in each panel were recorded with different detection methods (direct charge digitization or single pulse counting, respectively) and/or different gain voltages of the MCP detector depending on the signal height. In addition, the signals of monomers and dimers were blanked from reaching the detector during acquisition of the pulse counting spectra in order to avoid detector saturation. The varying detection probabilities connected with the detector settings are matched by a spectral overlap including at least one cluster mass peak. It is seen that in both cases sputtered neutral clusters containing up to more than 12 atoms can be observed. While the spectrum obtained under Xeþ bombardment is relatively clean, small peaks corresponding to Agn F and Agn S clusters are observed under SFþ 5 bombardment which are caused by a projectile induced contamination at the surface. The magnitude of these signals is negligible for sputtered neutrals but quite strong in the secondary ion
Fig. 1. Mass spectra of post-ionized neutral atoms and clusters sputtered from a polycrystalline silver surface under bombardment with 10-keV (a) Xeþ and (b) SFþ 5 ions incident under 45° with respect to the surface normal. Post-ionization laser: 193 nm, 5 107 W cm2 .
spectrum. Both spectra depicted in Fig. 1 have been normalized to the primary ion current. Due to the fact that the total signal heights observed in both spectra are comparable, it is evident that the respective partial sputtering yields do not change significantly between Xeþ and SFþ 5 projectiles. For a more quantitative evaluation, Fig. 2 shows the integrated signals, again normalized to the primary ion current, as a function of the number of atoms in the sputtered cluster. It is apparent that the partial sputtering yields of silver atoms and dimers sputtered from a polycrystalline silver surface are enhanced by factors of 1.5 and 1.2, respectively, while the yields of larger clusters remain practically unchanged upon switching from Xeþ to SFþ 5 . This observation clearly indicates that the total sputtering yield under SFþ 5 bombardment is comparable to that induced by Xeþ bombardment.
46
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48
Fig. 2. Integrated signal of neutral Agn clusters sputtered from a polycrystalline silver surface under bombardment with 10-keV Xeþ and SFþ 5 ions versus cluster size n. The data have been measured using setup 2 and normalized to the primary ion current.
It also coincides with what would be expected from a linear addition of the yields arising from the constituent atoms of the SF5 projectile impinging with the same velocity (using literature data [17] measured for Ne and Ar instead of F and S projectiles, respectively), thus indicating that any non-linear enhancement of the total sputtering yield under SFþ 5 bombardment is small. Fig. 3 shows the secondary ion signals of Agþ ions and Agþ n cluster ions sputtered under 11-keV Arþ , Xeþ and SFþ 5 which have been measured using setup 1. Since the measured signals have been normalized to the total primary ion current, the difference between the curves directly represents the difference in secondary ion yields induced by the different projectiles. The first important observation is that the secondary ions exhibit a much more pronounced yield increase than the corresponding neutral species. For the transition from Arþ to SFþ 5 projectiles, for instance, the average enhancement factor observed in Fig. 3 amounts to more than two orders of magnitude. The transition from Xeþ to SFþ 5 , being less pronounced, is quantitatively reproduced by similar data taken with setup 2, a finding which is reassuring in view of the largely different mass spectrometric techniques employed in both setups.
Fig. 3. Integrated signal of positive secondary atomic and cluster ions sputtered from a polycrystalline silver surface under bombardment with 11-keV Arþ , Xeþ and SFþ 5 ions. The data have been measured using setup 1 and normalized to the primary ion current.
From the comparison with the corresponding neutral data, it is immediately evident that the strong yield enhancement observed for secondary ions cannot be due to increased partial sputtering yields, but must be caused by a more efficient ionization of the sputtered atoms and clusters þ bombardunder SFþ 5 bombardment. Under Xe ment, the secondary ion signals are significantly below those of post-ionized neutrals, thus indicating a relatively low ionization probability which is typical for rare gas projectiles impinging onto a relatively clean metallic sample. It should be noted again that although the neutral and ion spectra depicted in Fig. 2 have been recorded under otherwise identical experimental conditions, the relative signal height does not necessarily reflect the ionization probability, since the post-ionization process leading to the detection of the neutral species is not saturated. This is particularly true for atoms and dimers which are ionized by a twophoton absorption process. Nevertheless, it is evident that the ionization probability is strongly enhanced under SFþ 5 bombardment. In view of the large matrix effects generally observed in SIMS, this finding is not surprising. It is known that the presence of electronegative species at the surface
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48
often significantly enhances the formation probability of positive secondary ions, an effect that is frequently called the ‘‘chemical enhancement effect’’ in SIMS [18]. Since the bombardment with SFþ 5 projectile ions will inevitably lead to a fluorine contamination at the surface, an enhancement of the ionization probability is therefore likely to be expected. The ionization probability of silver atoms and clusters sputtered under rare gas projectile bombardment has been determined to vary from around 104 for monomers to several 102 for large clusters ðn > 5Þ [19]. The data displayed in Fig. 2, which are in qualitative agreement with this finding, now indicate that a sizeable fraction of the clusters sputtered under SFþ 5 bombardment leaves the surface as positive secondary ions. A second important observation concerns the relative intensities of clusters with respect to that of the monomer. From Figs. 2 and 3, it is seen that – apart from the overall yield increase discussed above, the relative yields of clusters are practically þ identical for SFþ 5 and Xe projectiles. For the case studied here, the relative abundance of sputtered clusters is evidently not enhanced when polyatomic projectiles are used instead of atomic ions. Many experiments on cluster formation in sputtering have indicated that the abundance pattern of sputtered clusters is correlated with the total sputtering yield in such a way that higher sputtering yields lead to an increased abundance of large clusters [4]. Since the total sputtering yield is not significantly changed, one would therefore expect comparable abundance distributions under Xeþ and SFþ 5 bombardment. At least for the bombarding conditions employed here, a nonlinear (or non-additive) enhancement of the formation of large clusters under bombardment with a polyatomic projectile as has been frequently reported in the literature [20–22] is clearly not observed.
4. Conclusion The experiments reveal that the bombardment of a clean silver surface with SFþ 5 projectile ions leads to a drastic increase of the ionization probability of sputtered Ag atoms and Agn clusters as
47
compared to rare gas ion bombardment at the same kinetic energy. The partial sputtering yields, on the other hand, which may be taken as representative of the collisional formation mechanisms of sputtered clusters, remain practically unchanged. More specifically, the mass distribution of sputtered particles does not change in favor of þ projectiles. large clusters between SFþ 5 and Xe One must therefore conclude that – at least in the impact energy region explored here – non-linear enhancement effects due to the polyatomic nature of the SFþ 5 projectile do not occur. In particular, a drastically increased abundance of large clusters that has been observed in other experiments using polyatomic projectiles seems to be absent for the impact of SFþ 5 onto silver.
Acknowledgements The authors are very much indebted to NATO for supporting this work in the frame of the collaborative linkage grant PST.CLG.976317. One of the authors (S.G.) gladly acknowledges financial support in the frame of a doctorate stipendium issued by the Egyptian government.
References [1] H.M. Urbassek, W.O. Hofer, Mat. Fys. Medd. Dan. Vidensk. Selsk. 43 (1993) 97. [2] S.R. Coon, W.F. Calaway, M.J. Pellin, Nucl. Instr. and Meth. B 90 (1994) 518. [3] Z. Ma, S.R. Coon, W.F. Calaway, M.J. Pellin, D.M. Gruen, E.I. Nagy-Felsobuki, J. Vac. Sci. Technol. A 12 (1994) 2425. [4] A. Wucher, M. Wahl, in: A. Benninghoven, B. Hagenhoff, H.W. Werner (Eds.), Secondary Ion Mass Spectrometry (SIMS X), Wiley & Sons, Chichester, 1997, p. 64. [5] S.R. Coon, W.F. Calaway, M.J. Pellin, J.M. White, Surf. Sci. 298 (1993) 161. [6] A.D. Appelhans, J.E. Delmore, D.A. Dahl, Anal. Chem. 59 (1987) 1685. [7] B. Hagenhoff, P.L. Cobben, C. Bendel, E. Niehuis, A. Benninghoven, in: G. Gillen, R. Lareau, J. Bennett, F. Stevie (Eds.), Secondary Ion Mass Spectrometry (SIMS XI), Wiley & Sons, Chichester, 1998, p. 585. [8] A. Schnieders, M. Schroeder, D. Stapel, H.F. Arlinghaus, A. Benninghoven, in: A. Benninghoven, P. Bertrand, H.N. Migeon, H.W. Werner (Eds.), Secondary Ion Mass
48
[9]
[10] [11] [12] [13] [14] [15] [16]
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48 Spectrometry (SIMS XII), Elsevier Science, Amsterdam, 2000, p. 263. H. Yamamoto, F. Esaka, H. Asaoka, in: A. Benninghoven, P. Bertrand, H.N. Migeon, H.W. Werner (Eds.), Secondary Ion Mass Spectrometry (SIMS XII), Elsevier Science, Amsterdam, 2000, p. 295. C. Staudt, R. Heinrich, A. Wucher, Nucl. Instr. and Meth. B 164–165 (2000) 677. A.D. Bekkerman, N.Kh. Dzhemilev, V.M. Rotstein, Surf. Interface Anal. 15 (1990) 587. A. Wucher, M. Wahl, H. Oechsner, Nucl. Instr. and Meth. B 82 (1993) 337. A. Wucher, M. Wahl, H. Oechsner, Nucl. Instr. and Meth. B 83 (1993) 73. M. Wahl, A. Wucher, Nucl. Instr. and Meth. B 94 (1994) 36. R. Heinrich, A. Wucher, Nucl. Instr. and Meth. B 140 (1998) 27. A. Wucher, R. Heinrich, C. Staudt, in: A. Benninghoven, P. Bertrand, H.N. Migeon, H.W. Werner (Eds.), Second-
[17] [18]
[19]
[20]
[21]
[22]
ary Ion Mass Spectrometry (SIMS XII), Elsevier Science, Amsterdam, 2000, p. 143. H.H. Andersen, H.L. Bay, in: R. Behrisch (Ed.), Sputtering by Particle Bombardment I, Springer, 1981, p. 145. M.L. Yu, in: A. Benninghoven, Y. Nihei, H. Shimizu, H.W. Werner (Eds.), Secondary Ion Mass Spectrometry (SIMS IX), Wiley & Sons, Chichester, 1993, p. 10. R. Heinrich, C. Staudt, M. Wahl, A. Wucher, in: A. Benninghoven, P. Bertrand, H.N. Migeon, H.W. Werner (Eds.), Secondary Ion Mass Spectrometry (SIMS XII), Elsevier Science, Amsterdam, 2000, p. 111. S.F. Belykh, B. Habets, U.Kh. Rasulev, A.V. Samartsev, L.V. Stroev, I.V. Veryovkin, Nucl. Instr. and Meth. B 164– 165 (2000) 809. S.F. Belykh, V.I. Matveev, U.Kh. Rasulev, A.V. Samartsev, I.V. Veryovkin, in: G. Gillen, R. Lareau, J. Bennett, F. Stevie (Eds.), Secondary Ion Mass Spectrometry (SIMS XI), Wiley & Sons, Chichester, 1998, p. 957. S.F. Belykh, I.S. Bitensky, D. Mullajanov, U.K. Rasulev, Nucl. Instr. and Meth. B 129 (1997) 451.
Formation of sputtered silver clusters under bombardment with SFþ 5 ions S. Ghalab a, C. Staudt a, S.E. Maksimov b, P. Mazarov a, V.I. Tugushev b, N.Kh. Dzhemilev b, A. Wucher a,* a
b
Institute of Experimental Physics, University of Essen, D-45117 Essen, Germany Cluster Physics Laboratory, Arifov Institute of Electronics, 700125 Tashkent, Uzbekistan Received 30 May 2002
Abstract þ þ The formation of Agn clusters and Agþ n cluster ions under bombardment of a silver surface with SF5 and Xe projectile ions was investigated experimentally. In order to obtain information about the relative abundance of clusters among the flux of sputtered particles independent of their charge state, mass spectra of both secondary ions and sputtered neutral particles were recorded. The neutral species were post-ionized prior to mass analysis by means of photo-ionization using an intense UV laser at a wavelength of 193 nm. It is found that measured Agþ n signals increase þ þ significantly if SFþ 5 projectiles are used instead of rare gas (Ar or Xe ) ions of the same kinetic impact energy. The signals of neutral Ag atoms and Agn clusters, on the other hand, exhibit only a relatively small increase, thus indicating that the enhancement observed for the secondary ions is predominantly caused by an increased ionization probability of sputtered particles under SFþ 5 bombardment rather than by enhanced partial sputtering yields. While the transition from Arþ to Xeþ projectiles leads to a drastic increase of the relative abundance of larger clusters in the spectrum, practically no such effect can be detected for the transition from Arþ to SFþ 5 . This finding shows that the use of polyatomic SF5 projectiles does not lead to a higher efficiency in producing sputtered clusters. Ó 2002 Elsevier Science B.V. All rights reserved.
PACS: 79.20.Rf; 36.40.)c
1. Introduction If a solid is bombarded with keV-ions, particles are released from the surface due to elastic collisions (‘‘sputtering’’). It has been known for a long time that the flux of sputtered particles contains –
*
Corresponding author. Tel.: +49-201-183-4141; fax: +49201-183-93-4141. E-mail address: [email protected] (A. Wucher).
besides atomic species – molecules and clusters which may be composed of up to many hundred atoms [1]. The formation of such clusters in sputtering represents an intriguing phenomenon, since it is not at all clear how these species survive the relatively violent collisional processes that lead to the ejection of atoms from the bombarded surface. In general, the typical abundance pattern of clusters observed in sputtering strongly decreases with increasing cluster size. The quantitative relation between the relative abundance and the number of
0168-583X/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 2 ) 0 1 3 6 1 - 7
44
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48
atoms in a sputtered cluster is generally found to closely follow a power law with a decay exponent d that depends on the sputtering conditions [2–4]. As a general trend, it is found that the relative abundance of large clusters increases (and, hence, d decreases) with increasing total sputtering yield (i.e. the average number of atoms released per primary projectile ion) [4,5]. From a number of mass spectrometric experiments on secondary ions, i.e. those sputtered particles that leave the surface in either a positively or negatively charged ionic state, it has been suggested that the use of polyatomic projectiles may significantly enhance the sputtering yield. One prominent example projectile that has been demonstrated to be particularly effective in secondary ion mass spectrometry (SIMS) of organic surface species is SFþ 5 . More specifically, it has been shown that particularly the signal measured for complex molecular ions may be drastically enhanced under SFþ 5 bombardment [6–8]. These observations suggest that the relative abundance of clusters among the sputtered flux may also be enhanced if the surface is bombarded with SFþ 5 ions as compared to rare gas ion projectiles of the same kinetic energy [9]. The present paper is intended to examine this question by investigating the influence of the projectile on the composition of the flux of particles sputtered from a clean metallic surface. As a first step, we use silver as a target material since it is known that rare gas ion bombardment of a silver surface produces relatively large amounts of Agn clusters [10]. Since it is not a priori clear whether the secondary ion yield enhancements observed in the literature are caused by an increased formation or an increased ionization of the ejected species in the course of the sputtering process, it is necessary to investigate both ionic and neutral clusters leaving the surface.
2. Experimental The experiments reported in this work have been performed in collaboration between two laboratories using two different experimental setups. Part of the experimental data was taken with a secondary ion mass spectrometer located at the
Arifov Institute at Tashkent (setup 1) which has been described in detail elsewhere [11]. In short, a double focusing arrangement consisting of a magnetic sector and an electrostatic prism is used to detect ionic clusters which are sputtered from a polycrystalline silver sample under bombardment with Arþ , Xeþ or SFþ 5 ions impinging under 45° with respect to the surface normal. The primary ions are generated by an electron impact ion source with axial symmetric magnetic field delivering total beam currents of about 0.6 lA for Arþ , Xeþ and SFþ 5 ions. In order to permit a rapid switching between different projectile ions, the source was operated with a gas mixture of Argon, Xenon and SF6 , and the desired projectile was selected by means of a Wien filter. The energy of the primary ions was kept constant at 11 keV. By changing the polarity of the ion optical potentials, both positive and negative cluster ions can in principle be detected, the present experiments, however, are restricted to positive secondary ions. The setup is mounted in an ultrahigh vacuum (UHV) chamber with a base pressure of about 109 mbar. During the experiments, the working pressure rises to about 2 107 mbar due to the operation of the gas ion source. The second part of the experiments were performed with a laser post-ionization reflectron time-of-flight mass spectrometer located at the University of Essen (setup 2, base pressure 109 mbar, working pressure 3 109 mbar) which has also been previously described [12–15]. In this setup, a polycrystalline silver sample is bombarded under 45° incidence with Xeþ or SFþ 5 ions of 10 keV. The ion beam is generated by a commercial cold cathode plasma ion source (Atomica Microfocus) that delivers total beam currents of 330 nA (Xeþ ) and 77 nA (SFþ 5 ) when operated with Xenon or SF6 gas under otherwise comparable conditions. During data acquisition, the primary ion beam was operated in a pulsed mode with a pulse length of 10 ls at a repetition rate of 10 Hz. Mass spectra of neutral particles sputtered from the surface are recorded by post-ionization of the ejected neutral species using an intense UV laser pulse. The ionizing radiation was generated by an excimer laser operated at k ¼ 193 nm with a cross-section of about 2.5 mm2 in the interaction region located
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48
45
about 1 mm above the surface. Secondary ion spectra are measured by simply switching the ionization laser off and leaving the remainder of the experiment unchanged. This way, secondary ions and neutrals are detected under otherwise the same experimental conditions [16]. The procedures employed for data acquisition with this instrument have been described in much detail elsewhere [16]. The recorded time-of-flight spectra were averaged over typically several 1000 instrument cycles, i.e. primary ion pulses and – where applicable – laser shots. Note that in this particular experiment the photo-ionization efficiency is not saturated, and the measured signals of post-ionized neutrals therefore strongly depend on the laser power density. The pulse energy of the laser was therefore carefully monitored and kept constant within a few percent by means of an optical feedback loop.
3. Results and discussion Fig. 1 shows mass spectra of post-ionized sputtered neutral atoms and clusters which have been ejected from the polycrystalline silver surface under bombardment with 10-keV Xeþ and SFþ 5 ions under otherwise identical experimental conditions (setup 2). The different traces depicted in each panel were recorded with different detection methods (direct charge digitization or single pulse counting, respectively) and/or different gain voltages of the MCP detector depending on the signal height. In addition, the signals of monomers and dimers were blanked from reaching the detector during acquisition of the pulse counting spectra in order to avoid detector saturation. The varying detection probabilities connected with the detector settings are matched by a spectral overlap including at least one cluster mass peak. It is seen that in both cases sputtered neutral clusters containing up to more than 12 atoms can be observed. While the spectrum obtained under Xeþ bombardment is relatively clean, small peaks corresponding to Agn F and Agn S clusters are observed under SFþ 5 bombardment which are caused by a projectile induced contamination at the surface. The magnitude of these signals is negligible for sputtered neutrals but quite strong in the secondary ion
Fig. 1. Mass spectra of post-ionized neutral atoms and clusters sputtered from a polycrystalline silver surface under bombardment with 10-keV (a) Xeþ and (b) SFþ 5 ions incident under 45° with respect to the surface normal. Post-ionization laser: 193 nm, 5 107 W cm2 .
spectrum. Both spectra depicted in Fig. 1 have been normalized to the primary ion current. Due to the fact that the total signal heights observed in both spectra are comparable, it is evident that the respective partial sputtering yields do not change significantly between Xeþ and SFþ 5 projectiles. For a more quantitative evaluation, Fig. 2 shows the integrated signals, again normalized to the primary ion current, as a function of the number of atoms in the sputtered cluster. It is apparent that the partial sputtering yields of silver atoms and dimers sputtered from a polycrystalline silver surface are enhanced by factors of 1.5 and 1.2, respectively, while the yields of larger clusters remain practically unchanged upon switching from Xeþ to SFþ 5 . This observation clearly indicates that the total sputtering yield under SFþ 5 bombardment is comparable to that induced by Xeþ bombardment.
46
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48
Fig. 2. Integrated signal of neutral Agn clusters sputtered from a polycrystalline silver surface under bombardment with 10-keV Xeþ and SFþ 5 ions versus cluster size n. The data have been measured using setup 2 and normalized to the primary ion current.
It also coincides with what would be expected from a linear addition of the yields arising from the constituent atoms of the SF5 projectile impinging with the same velocity (using literature data [17] measured for Ne and Ar instead of F and S projectiles, respectively), thus indicating that any non-linear enhancement of the total sputtering yield under SFþ 5 bombardment is small. Fig. 3 shows the secondary ion signals of Agþ ions and Agþ n cluster ions sputtered under 11-keV Arþ , Xeþ and SFþ 5 which have been measured using setup 1. Since the measured signals have been normalized to the total primary ion current, the difference between the curves directly represents the difference in secondary ion yields induced by the different projectiles. The first important observation is that the secondary ions exhibit a much more pronounced yield increase than the corresponding neutral species. For the transition from Arþ to SFþ 5 projectiles, for instance, the average enhancement factor observed in Fig. 3 amounts to more than two orders of magnitude. The transition from Xeþ to SFþ 5 , being less pronounced, is quantitatively reproduced by similar data taken with setup 2, a finding which is reassuring in view of the largely different mass spectrometric techniques employed in both setups.
Fig. 3. Integrated signal of positive secondary atomic and cluster ions sputtered from a polycrystalline silver surface under bombardment with 11-keV Arþ , Xeþ and SFþ 5 ions. The data have been measured using setup 1 and normalized to the primary ion current.
From the comparison with the corresponding neutral data, it is immediately evident that the strong yield enhancement observed for secondary ions cannot be due to increased partial sputtering yields, but must be caused by a more efficient ionization of the sputtered atoms and clusters þ bombardunder SFþ 5 bombardment. Under Xe ment, the secondary ion signals are significantly below those of post-ionized neutrals, thus indicating a relatively low ionization probability which is typical for rare gas projectiles impinging onto a relatively clean metallic sample. It should be noted again that although the neutral and ion spectra depicted in Fig. 2 have been recorded under otherwise identical experimental conditions, the relative signal height does not necessarily reflect the ionization probability, since the post-ionization process leading to the detection of the neutral species is not saturated. This is particularly true for atoms and dimers which are ionized by a twophoton absorption process. Nevertheless, it is evident that the ionization probability is strongly enhanced under SFþ 5 bombardment. In view of the large matrix effects generally observed in SIMS, this finding is not surprising. It is known that the presence of electronegative species at the surface
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48
often significantly enhances the formation probability of positive secondary ions, an effect that is frequently called the ‘‘chemical enhancement effect’’ in SIMS [18]. Since the bombardment with SFþ 5 projectile ions will inevitably lead to a fluorine contamination at the surface, an enhancement of the ionization probability is therefore likely to be expected. The ionization probability of silver atoms and clusters sputtered under rare gas projectile bombardment has been determined to vary from around 104 for monomers to several 102 for large clusters ðn > 5Þ [19]. The data displayed in Fig. 2, which are in qualitative agreement with this finding, now indicate that a sizeable fraction of the clusters sputtered under SFþ 5 bombardment leaves the surface as positive secondary ions. A second important observation concerns the relative intensities of clusters with respect to that of the monomer. From Figs. 2 and 3, it is seen that – apart from the overall yield increase discussed above, the relative yields of clusters are practically þ identical for SFþ 5 and Xe projectiles. For the case studied here, the relative abundance of sputtered clusters is evidently not enhanced when polyatomic projectiles are used instead of atomic ions. Many experiments on cluster formation in sputtering have indicated that the abundance pattern of sputtered clusters is correlated with the total sputtering yield in such a way that higher sputtering yields lead to an increased abundance of large clusters [4]. Since the total sputtering yield is not significantly changed, one would therefore expect comparable abundance distributions under Xeþ and SFþ 5 bombardment. At least for the bombarding conditions employed here, a nonlinear (or non-additive) enhancement of the formation of large clusters under bombardment with a polyatomic projectile as has been frequently reported in the literature [20–22] is clearly not observed.
4. Conclusion The experiments reveal that the bombardment of a clean silver surface with SFþ 5 projectile ions leads to a drastic increase of the ionization probability of sputtered Ag atoms and Agn clusters as
47
compared to rare gas ion bombardment at the same kinetic energy. The partial sputtering yields, on the other hand, which may be taken as representative of the collisional formation mechanisms of sputtered clusters, remain practically unchanged. More specifically, the mass distribution of sputtered particles does not change in favor of þ projectiles. large clusters between SFþ 5 and Xe One must therefore conclude that – at least in the impact energy region explored here – non-linear enhancement effects due to the polyatomic nature of the SFþ 5 projectile do not occur. In particular, a drastically increased abundance of large clusters that has been observed in other experiments using polyatomic projectiles seems to be absent for the impact of SFþ 5 onto silver.
Acknowledgements The authors are very much indebted to NATO for supporting this work in the frame of the collaborative linkage grant PST.CLG.976317. One of the authors (S.G.) gladly acknowledges financial support in the frame of a doctorate stipendium issued by the Egyptian government.
References [1] H.M. Urbassek, W.O. Hofer, Mat. Fys. Medd. Dan. Vidensk. Selsk. 43 (1993) 97. [2] S.R. Coon, W.F. Calaway, M.J. Pellin, Nucl. Instr. and Meth. B 90 (1994) 518. [3] Z. Ma, S.R. Coon, W.F. Calaway, M.J. Pellin, D.M. Gruen, E.I. Nagy-Felsobuki, J. Vac. Sci. Technol. A 12 (1994) 2425. [4] A. Wucher, M. Wahl, in: A. Benninghoven, B. Hagenhoff, H.W. Werner (Eds.), Secondary Ion Mass Spectrometry (SIMS X), Wiley & Sons, Chichester, 1997, p. 64. [5] S.R. Coon, W.F. Calaway, M.J. Pellin, J.M. White, Surf. Sci. 298 (1993) 161. [6] A.D. Appelhans, J.E. Delmore, D.A. Dahl, Anal. Chem. 59 (1987) 1685. [7] B. Hagenhoff, P.L. Cobben, C. Bendel, E. Niehuis, A. Benninghoven, in: G. Gillen, R. Lareau, J. Bennett, F. Stevie (Eds.), Secondary Ion Mass Spectrometry (SIMS XI), Wiley & Sons, Chichester, 1998, p. 585. [8] A. Schnieders, M. Schroeder, D. Stapel, H.F. Arlinghaus, A. Benninghoven, in: A. Benninghoven, P. Bertrand, H.N. Migeon, H.W. Werner (Eds.), Secondary Ion Mass
48
[9]
[10] [11] [12] [13] [14] [15] [16]
S. Ghalab et al. / Nucl. Instr. and Meth. in Phys. Res. B 197 (2002) 43–48 Spectrometry (SIMS XII), Elsevier Science, Amsterdam, 2000, p. 263. H. Yamamoto, F. Esaka, H. Asaoka, in: A. Benninghoven, P. Bertrand, H.N. Migeon, H.W. Werner (Eds.), Secondary Ion Mass Spectrometry (SIMS XII), Elsevier Science, Amsterdam, 2000, p. 295. C. Staudt, R. Heinrich, A. Wucher, Nucl. Instr. and Meth. B 164–165 (2000) 677. A.D. Bekkerman, N.Kh. Dzhemilev, V.M. Rotstein, Surf. Interface Anal. 15 (1990) 587. A. Wucher, M. Wahl, H. Oechsner, Nucl. Instr. and Meth. B 82 (1993) 337. A. Wucher, M. Wahl, H. Oechsner, Nucl. Instr. and Meth. B 83 (1993) 73. M. Wahl, A. Wucher, Nucl. Instr. and Meth. B 94 (1994) 36. R. Heinrich, A. Wucher, Nucl. Instr. and Meth. B 140 (1998) 27. A. Wucher, R. Heinrich, C. Staudt, in: A. Benninghoven, P. Bertrand, H.N. Migeon, H.W. Werner (Eds.), Second-
[17] [18]
[19]
[20]
[21]
[22]
ary Ion Mass Spectrometry (SIMS XII), Elsevier Science, Amsterdam, 2000, p. 143. H.H. Andersen, H.L. Bay, in: R. Behrisch (Ed.), Sputtering by Particle Bombardment I, Springer, 1981, p. 145. M.L. Yu, in: A. Benninghoven, Y. Nihei, H. Shimizu, H.W. Werner (Eds.), Secondary Ion Mass Spectrometry (SIMS IX), Wiley & Sons, Chichester, 1993, p. 10. R. Heinrich, C. Staudt, M. Wahl, A. Wucher, in: A. Benninghoven, P. Bertrand, H.N. Migeon, H.W. Werner (Eds.), Secondary Ion Mass Spectrometry (SIMS XII), Elsevier Science, Amsterdam, 2000, p. 111. S.F. Belykh, B. Habets, U.Kh. Rasulev, A.V. Samartsev, L.V. Stroev, I.V. Veryovkin, Nucl. Instr. and Meth. B 164– 165 (2000) 809. S.F. Belykh, V.I. Matveev, U.Kh. Rasulev, A.V. Samartsev, I.V. Veryovkin, in: G. Gillen, R. Lareau, J. Bennett, F. Stevie (Eds.), Secondary Ion Mass Spectrometry (SIMS XI), Wiley & Sons, Chichester, 1998, p. 957. S.F. Belykh, I.S. Bitensky, D. Mullajanov, U.K. Rasulev, Nucl. Instr. and Meth. B 129 (1997) 451.