Документ взят из кэша поисковой машины. Адрес оригинального документа : http://danp.sinp.msu.ru/LIIWM/Nimb249_06_65Khodyrev.pdf Дата изменения: Sun Oct 21 20:50:32 2007 Дата индексирования: Mon Oct 1 22:15:05 2012 Кодировка: ISO8859-5

N IM B
Beam Interactions with Materials & Atoms

Nuclear Instruments and Methods in Physics Research B 249 (2006) 65-68 www.elsevier.com/locate/nimb

The Coulomb explosion of swift C? molecules under 2 channeling conditions
Д R. Gonzalez-Arrabal
a

a,*

Д , V.A. Khodyrev b, N. Gordillo a, G. Garcia a, D.O. Boerma

a

Д CMAM, Universidad Autonoma de Madrid, C/Faraday 3, Cantoblanco, 28049 Madrid, Spain b Institute of Nuclear Physics, Moscow State University, Moscow 119899, Russia Available online 16 May 2006

Abstract We have bombarded crystalline Si with high-energy (1.847 MeV/atom) atomic C+ and molecular C? ions in h111i, h10 0i and h211i 2 axial channeling and in random directions to observe the effect of Coulomb explosion (CE) for the molecular ions. The energy spectra of backscattered ions measured with equivalent fluences of atomic and molecular ions are compared. In channeling spectra measured with C+ and C? beams the significant de-channeling due to damage build-up caused by the irradiation was found to be equal for atoms and 2 molecules and linearly increasing with fluence in the studied range. In addition to the damage effect the CE effect was clearly observed for the three channels studied. г 2006 Published by Elsevier B.V.
PACS: 79.20.Rf; 34.10.+x; 34.50.Dy; 34.50.Ez Keywords: Coulomb explosion; Molecular beams; Ion-solid interaction; Radiation damage

1. Introduction Beams of swift molecular and multi-atom cluster ions are a relatively new tool to locally modify the structure of thin layers, thereby altering the electronic, magnetic or optical properties on the nanometer scale. For instance, molecular beams are used for implantation in silicon [1] and single ion track formation with a C60 molecular beam has been investigated [2]. A main difference with implantation of atomic ions is the fact that the atoms in the molecule travel at a relatively short distance from each other, so that correlated damage may be produced. This has as a consequence that the energy loss, the average charge state and the damage formation may not be simply a linear superposition of these quantities determined for the atomic constituents of the molecules [3].

Corresponding author. Tel.: +34 91 497 24 22; fax: +34 497 36 23. Д E-mail address: raquelgonzalezarrabal@uam.es (R. Gonzalez-Arrabal). 0168-583X/$ - see front matter г 2006 Published by Elsevier B.V. doi:10.1016/j.nimb.2006.03.080

*

When molecules enter into a solid, the atoms will be stripped of some electrons in the first few surface layers. In the case of a di-atomic molecule a Coulomb energy is acquired which is equal to q1 Ц q2/D, where q1, q2 are the charges on the atoms forming the molecule and D is the inter-atomic distance. In the following Coulomb explosion (CE) this potential energy is gradually transferred into kinetic energy. The CE of the charged molecular fragments has been extensively studied [4-9] by measuring the velocity vector and the charge of the fragments behind a thin carbon foil used as a target. The structure of small molecules can be accurately determined from such measurements. From these works it can be concluded that the average charge state of the atomic fragments of small molecules (N2, C2, C3 with an energy of $2 MeV/atom) measured behind a thin foil is slightly smaller (by a factor of $0.9) than the average charge of $3+ of atoms passing the foil with the same energy [8,9]. The effect of the CE on the channeling of H2, H3 molecular ions with energies in the 0.15-2.4 MeV/atom range has been studied in crystalline silicon [10-12]. A clear increase of the de-channeling rate


66

Д R. Gonzalez-Arrabal et al. / Nucl. Instr. and Meth. in Phys. Res. B 249 (2006) 65-68

for molecular ions as compared to atomic ions was observed. The results for 0.65 MeV/atom were reproduced in computer simulations [12], where in the calculation of the trajectories the mutual interaction of the charged fragments was taken into account. Here an average charge smaller than unity on the molecular fragments had to be assumed to get a good fit. In this work we study the de-channeling of 1.847 MeV C+ ions and the fragments of exploding C? molecules of 2 the same velocity in a number of axial channels of a Si single crystal. The final aim of this work is to simulate the data with the same program as used for the simulation of the de-channeling of H-molecules [12], where the interaction of the molecule fragments is taken into account. In a sense, the molecule fragments are here spectators of each others charge state. By comparing measured and simulated results, information on the average charge state of the fragments could be obtained. Here we describe the measurement procedure and show that the CE effect can be measured. The comparison with simulations will be the subject of a future paper. To study the de-channeling due to the CE effect also the radiation-induced damage by the channeling atomic and molecular ions should be measured. This part of the present study is interesting in itself, because of the possibility to observe effects of correlated damage formation by the molecular fragments. Such studies have been previously done in random orientation. However, measuring in channeling orientation offers advantages in comparison with measurements done in a random direction. For example, if the measurements are performed in channeling orientation two regimes can be considered. The first regime corresponds to a thin surface layer (corresponding to the surface peak visible in channeling spectra) which is subjected to bombardment with the full flux of molecular fragments still traveling at a short distance from each other. Assuming an А average charge state of 3+ and an initial distance of 1.54 A А this distance increases by only 20% over the first 70 A. The second regime is associated with deeper layers in which the close ion-atom collisions are much less probable due to the redistribution of the ion flux and where the fragments move at a larger distance from each other. In this publication we will focus on the study of the CE effect within the bulk region. 2. Experimental details The measurements were performed in a vacuum chamber equipped with a four axis goniometer, which is connected to the 5 MV tandetron accelerator at the CMAM/ UAM [13]. The spectra of scattered C atoms were measured with atomic (C+) and molecular (C? ) ions with an 2 energy of 1.847 MeV/atom. A Cs-sputter source is used (HVE model 860A) at the low energy side of the accelerator with a cathode consisting of graphite powder hammered into a copper housing. When the pressure of the N2 stripper gas in the terminal is small enough, C? 2

ions are produced in the stripper canal, with currents of the order of 10 nA, suitable for analysis or modification of materials. The species of the ions emerging from the accelerator was selected by the analysing magnet and it was further checked that in RBS measurements spectra taken with molecular or atomic beams with the same velocity and atom fluence are identical in random configuration. This confirms that the beam analysed by the switching magnet after acceleration is indeed C? . The beam line has two sets of collimator slits 2 (0.7 and 2.7 m upstream from the sample). The measurements were performed in the h111i, h10 0i and h211i channeling orientations. The slit apertures were set at 2.0 and 2.0 mm for the h100i orientation, and 1.5 and 1.0 mm for h21 1i and h11 1i orientations of the crystal. A H-terminated Si(1 0 0) crystal obtained by chemically etching with dilute HF was chosen to study the CE effect. The scattered particles were detected by a silicon surface-barrier detector with an energy resolution of 45 keV for our conditions. No degradation of the detector resolution or calibration due to radiation damage was observed during the measurements. For the ion- crystal combination considered we had to optimize the scattering geometry to compensate for the limited energy resolution of the detector, combined with a small value of the kinematic factor k. For scattering close to the backward direction the latter is equal to 0.16. With this geometry we would obtain the whole spectrum in an energy range comparable with the detector resolution. Therefore, the position of the detector was chosen to be at a scattering angle of 120А for the h10 0i orientation, and at 80А for the h11 1i and h21 1i orientations, so that the kinematic factor is equal to 0.25 and 0.47, respectively. The influence of the CE effect on the shape and yield of the channeling spectra was found to be comparable with the modification of the spectra due to the crystal damage induced by the irradiation necessary for taking the spectra. Therefore, to measure the small CE effect in a reliable way, the following measures were taken: (1) To ensure that equivalent numbers of current integrator counts correspond to the same number of atoms for the atomic or the molecular beams, an electron suppressor ring biased at Р180 V was placed at the entrance of the chamber. A bias voltage of +160 V was applied to the sample holder to avoid a possible contribution of the secondary electrons to the current integration. C-backscattered spectra were measured in random configuration near to each channeling orientation with an atomic and a molecular beam to test that for equivalent numbers of current integrator counts the spectra taken with atomic and molecular beams were the same. This was indeed the case. The beam conditions were chosen such that the size of the beam on the sample was determined by the collimation of the slits, and not by focussing. Under these conditions we assume that the same spot of the crystal is irradiated homogeneously, both for atomic and molecular beams. C-back-


Д R. Gonzalez-Arrabal et al. / Nucl. Instr. and Meth. in Phys. Res. B 249 (2006) 65-68

67

scattered spectra for each channeling orientation were sequentially measured on the same spot of the sample with equal atom fluences, alternating irradiation with atoms and molecules to obtain the necessary information on the damage formation both by atoms and by molecules. In this way we avoided to move the goniometer between measurements, thereby eliminating possible deviations on the channeling alignment between measurements with atomic and molecular beams, respectively. (2) Dead-time correction of the spectra was implied by connecting a pulse generator triggered by pulses from the current integrator to the test input of the preamplifier. The number of counts in the test peak in the spectra thus obtained was used to renormalize the spectra. The dead time was only a few percent and comparable for spectra taken with atoms or molecules. 3. Results and discussion C-backscattered spectra for different channeling orientations, taken with equal atom fluences, are shown in Fig. 1. For all orientations, the spectra with the lowest yield were taken on a virgin spot with an atomic beam. The next spec-

tra in the sequence of increasing yield are taken with a molecular beam, and the next two with atomic beams. It can be observed that in all channeling directions the yield for equal fluences increases due to de-channeling after each irradiation step. To study in more detail the changes observed in the spectra for different axial channels we will consider further the yield in two windows, B (bulk) and S (surface), indicated in Fig. 1. Fig. 2 shows the yield in window B as a function of the irradiation step (and thus as a function of the fluence) for different channeling orientations. The yields are normalized to the yield on the virgin spectrum for each channel; therefore, the yield for the first irradiation with atoms is equal to unity. The statistical errors are smaller than the symbol size. The data for the atomic beam displays a linear increase of the yield with fluence due to the crystal damage in the studied range in all the channeling orientations. The data points taken with a molecular beam deviate from this trend: these points reveal a higher dechanneling for molecular beams. This extra yield is not due to extra damage formation, because in the next spectrum taken with atoms the yield is in line with the linear increase in the yield. Thus, since the correlation effect in

<111>
100 8

Depth (nm)
50 0

<100>
200

Depth (nm)
100 0

<211>
C + C2
+

Depth (nm)
150 100 50 0

200 5 4

Yield (x103)

B
6

C + C2 C + C
+

+

12 10

B S

Yield (x103)

S
4 2

8 6 4 2

Yield (x103)

C + C

+

C + C2

+

B
3 2 1 0

C + C

+

S

0

0

50

75

100

125

20

30

40

50

60

70

25

50

75

Channel

Channel

Channel

Fig. 1. C-backscattered spectra for h111i, h10 0i and h211i channeling orientation measured with fluences of 1.4 З 1015 cmР2 for the h111i and h211i 4.3 З 1014 cmР2 for h100i with atomic (C+) and molecular (C? ) beams. The irradiation sequence for the h111i and h211i channels was C+ (black lines), 2 C? (open circles), C+ (full squares), C+ (crosses). In all frames the height of the spectra increases monotonically with the fluence. 2

<111> B
1.20 1.5

<100> B C
Normalized area
1.15
+

1.4

<211> B C
+ +

C

+ +

Normalized area

1.4 1.3 1.2 1.1 1.0 1

C2

Normalized area

+ C2

1.3

C2
CE

1.10

CE

1.2

CE

1.05

1.1

1.00 2 3 4 1 2 3 4

1.0 1 2 3 4

Irradiation sequence

Irradiation sequence

Irradiation sequence

Fig. 2. Normalized yield to that of the virgin spectrum for the h111i, h100i and h211i channeling directions within the bulk windows B as a function of the irradiation step. The fluence per step was 1.4 З 1015 cmР2 for the h111i and h211i orientations and 4.3 З 1014 cmР2 for the h100i orientation. The square dots indicate the yield measured with atomic C+ beams; the stars indicate the yields measured with molecular C? beams. The lines just connect the 2 first and last points.


68

Д R. Gonzalez-Arrabal et al. / Nucl. Instr. and Meth. in Phys. Res. B 249 (2006) 65-68

the damage formation can be neglected in the bulk region, the extra de-channeling observed for irradiation with molecules must be ascribed to the increase of the transverse energy due to the CE. By comparing the enhancement factor of the yield after irradiation with a molecular beam for each channeling direction, some differences are observed. In our particular case these differences are partly due to the different depth scaling in the sample measured for different geometries as shown in Fig. 1. Note that the depth corresponding to window B is narrower by roughly a factor of 2 for the h111i channel as compared to the other two channels. 4. Conclusions The damage formation in the bulk region was found to be equal for atoms and molecules, indicating that no correlated damage is produced within this region. The CE effect was clearly observed in addition to a significant increase in de-channeling due to damage formation for different channels.

References
[1] J.A. Davies, G. Foti, L.M. Howe, J.B. Mitchell, K.B. Winterbon, Phys. Rev. Lett. 34 (1975) 1441. [2] B. Canut, S.M.M. Ramos, Radiat. Eff. Defects Solids 145 (1997) 1. [3] P. Sigmund, I.S. Bitensky, J. Jensen, Nucl. Instr. and Meth. B 112 (1996) 1. [4] Z. Vager, R. Naaman, E.P. Kanter, Sci. New Ser. 244 (1989) 426. [5] W. Brandt, R.H. Ritchie, Nucl. Instr. and Meth. B 132 (1976) 43. [6] R. Wester, F. Albrecht, M. Grieser, L. Knoll, R. Repnow, D. Schwalm, A. Wolf, A. Baer, J. Levin, Z. Vager, D. Zajfman, Nucl. Instr. and Meth. A 413 (1998) 379. [7] D. Maor, P.J. Cooney, A. Faibis, E.P. Kanter, W. Koenig, B.J. Zabransky, Phys. Rev. A 32 (1985) 105. [8] I. Plesser, Z. Vager, R. Nasman, Phys. Rev. Lett. 56 (1986) 1559. [9] A. Brunelle, S. Della-Negra, J. Depauw, D. Jacquet, Y. Le Beyec, M. Pautrat, Phys. Rev. A 59 (1999) 4456. [10] T.A. Tombrello, J.M. Caywood, Phys. Rev. B 8 (1973) 3065. [11] R.C. Fadanelli, P.L. Grande, M. Behar, F. Dias, G. Schiwietz, C.D. Denton, Phys. Rev. B 69 (2004) 212104. [12] V.A. Khodyrev, V.S. Kulikauskas, C. Yang, Nucl. Instr. and Meth. B 195 (2002) 259. [13] D.J.W. Mous, A. Gottdang, R.G. Haitsma, G. Garcia Lopez, ДД A. Climent-Font, F. Agullo-Lopez, D.O. Boerma, in: J.L. Duggan, I.L. Morgan (Eds.), CP680, Application of Accelerators in Research and Industry: 17th Int. Conf., Am. Inst. of Physics, 2003, p. 999.