1: \documentstyle[aps,floats,graphicx]{revtex}
2: %\usepackage[dvips]{graphics,monochrome}
3:
4: \begin{document}
5:
6: \newcommand{\bec}{\begin{center}}
7: \newcommand{\ec}{\end{center}}
8: \newcommand{\be}{\begin{equation}}
9: \newcommand{\ee}{\end{equation}}
10: \newcommand{\beqn}{\begin{eqnarray}}
11: \newcommand{\eeqn}{\end{eqnarray}}
12: \newcommand{\bet}{\begin{table}}
13: \newcommand{\ent}{\end{table}}
14: \newcommand{\bib}{\bibitem}
15:
16:
17: %\baselineskip 4.2mm
18:
19:
20: \wideabs{
21:
22: \title{
23: Fingerprint of super-interdiffusion: anomalous intermixing in Pt/Ti
24: %Interface-anisotropy induced anomalous intermixing in Pt/Ti
25: %Interface-anisotropy induced asymmetry of intermixing in bilayers
26: }
27:
28:
29: \author{P. S\"ule, M. Menyh\'ard}
30: \address{Research Institute for Technical Physics and Material Science,
31: www.mfa.kfki.hu/$\sim$sule, sule@mfa.kfki.hu\\
32: Konkoly Thege u. 29-33, Budapest, Hungary\\
33: %$^{\star}$ National Institute of Standards \& Technology, Gaithersburg, Maryland 20899\\
34: }
35: %\email{sule@mfa.kfki.hu}
36:
37: \date{\today}
38:
39: \maketitle
40:
41: \begin{abstract}
42: The ion-sputtering induced transient enhanced intermixing has been studied by molecular dynamics (MD) simulations
43: in Pt/Ti and its anomalous nature has been explained as a superdiffusive transient enhanced interdiffusion.
44: We find ballistic mixing and a robust mass effect in Pt/Ti. The sum of the square of atomic displacements
45: ($\langle R^2 \rangle$)
46: asymptotically grows nonlinearily and scales as $N^2$ and $\sim t^2$, where $N$ and $t$ are the ion-number fluence and the time of ion-sputtering, respectively.
47: This anomalous behavior explains the high diffusity tail in the concentration profile obtained
48: by
49: Auger electron spectroscopy depth profiling (AES-DP) analysis in Pt/Ti/Si substrate (Pt/Ti) multilayer.
50: In Ta/Ti/Pt/Si multilayer we find a linear time scaling of $\langle R^2 \rangle \propto t$ at the Ti/Pt interface indicating
51: the suppression of superdiffusive features.
52: We propose a qualitative explanation based on the accelerative effect of nonlinear forces provbided by the
53: anharmonic host lattice.
54:
55:
56: {\em PACS numbers:} 66.30.Jt, 61.80.Jh, 68.35.-p, 68.35.Fx, 66.30.-h, 68.55.-a \\
57: {\scriptsize {\em Keywords:} intermixing, interdiffusion, anomalous atomic transport, interface, impurity diffusion, Auger depth profiling, sputtering, computer simulations, multilayer, ion-solid interaction, molecular dynamics, Ti/Pt, accelerative effects
58: }
59: \end{abstract}
60: }
61:
62:
63: \section{Introduction}
64:
65:
66: The most well known atomic diffusion mechanisms (simple site exchange, vacancy, self-interstitial mediated or concentration dependent) are thermally activated
67: processes \cite{Michely}.
68: However, there are a growing number of evidences are emerged
69: that the anomalous broadening of interfaces or the high diffusity tail in the impurity concentration profile
70: could be the fingerprint of anomalously fast impurity diffusion
71: \cite{Abrasonis,Buchanan,Parascandola}.
72: It has been concluded by Abrasonis {\em et al.} that anomalously long diffusion depths
73: can only be understood by the assistance of lattice accelerative effects \cite{Abrasonis}.
74: %Also, transient inter-layer atomic mobility could occur in few systems
75: %\cite{Sule_condmat}.
76: Transient enhanced diffusion has also been reported in post-annealed
77: dopant implanted semiconductors \cite{Delugas,Melis,Solmi,Sparks,Venezia,Slijkerman}.
78:
79:
80: Anomalously long interdiffusion dephts has been found in various diffusion
81: couples \cite{Abrasonis,Buchanan,Delugas,Melis,Solmi,Sparks,Venezia,Slijkerman}.
82: % The anomalous nature of these atomic transport processes manifests in various ways.
83: It has been reported that
84: anomalous interdiffusion is not driven by bulk diffusion parameters nor by thermodynamic forces (such as heats of alloying) \cite{Buchanan} or nor by heats of mixing \cite{Sule_NIMB04,Sule_PRB05}.
85: Transient intermixing (IM, surface alloying) has also been found the most recently by molecular dynamics simulations at cryogenic temperatures
86: during the deposition of Pt on Al(111)
87: \cite{Sule_condmat}.
88:
89:
90:
91:
92: Computer simulations have revealed that
93: the mass-anisotropy of metallic bilayers governs ion-bombardment induced enhancement of interdiffusion at the interface \cite{Sule_PRB05} and
94: greatly influences surface morphology development
95: \cite{Sule_SUCI,Sule_NIMB04_2} and intermixing properties during ion-sputtering \cite{Sule_NIMB04,Sule_PRB05}.
96: %The heavy atoms at the interface block the transport path
97: %of the overlayer atoms temporarily and reverse the flux of energetic light atoms towards the surface which could lead to mixing induced surface roughening \cite{Sule_SUCI,Sule_NIMB04_2}
98: %and to the amplification of atomic transport in the bulk \cite{Sule_PRB05}.
99: In mass-anisotropic metallic bilayers, such e.g. Al/Pt, or in Ti/Pt low-energy ion-bombardment leads also to
100: anomalously large broadening at the interface \cite{Sule_NIMB04,Sule_PRB05,Sule_SUCI}.
101: The intermixing length (and the mixing efficiency) scales nonlinearily with the mass anisotropy (mass ratio)
102: \cite{Sule_PRB05}.
103: These enhanced atomistic transport processes cannot be understood by the established mechanisms of radiation-enhanced
104: diffusion (RED) \cite{Abrasonis}.
105: This is because RED cannot cause anomalously large diffusion depths (far beyond the ion penetration depth) \cite{Abrasonis}.
106: Moreover, the observed asymmetry of intermixing in Ti/Pt and in Pt/Ti can not be understood
107: by the thermal spike model nor by RED \cite{Sule_condmat06b}.
108: Also, one might not explain the observed large intermixing length \cite{Buchanan} and
109: abrupt transport processes \cite{Sule_condmat} by
110: thermally activated diffusion.
111:
112: It has already been shown recently that the asymmetry of intermixing has been found
113: in Pt/Ti and in Ti/Pt bilayers. However, no explanation is found for the very strong
114: and much weaker interdiffusion in Pt/Ti and in Ti/Pt, respectively \cite{Sule_condmat06b}.
115: We propose in this article to understand the enhanced interdiffusion in Pt/Ti as a superdiffusive atomic transport
116: process
117: since they fulfill the most important condition of superdiffusion:
118: the square of atomic displacements scales nonlinearily with the time of diffusion \cite{anomalous}.
119: We present evidences in this paper for the fullfillment of this condition for Pt/Ti.
120: We would like to show that ballistic (athermal) intermixing
121: occurs in Pt/Ti in contrast to Ti/Pt in which normal
122: radiation-enhanced thermally activated interdiffusion takes place.
123: %An interpretation is attempted on the basis of
124: %interface-anisotropy induced enhancement of atomic mobility which
125: %manifests as a long-range diffusity tail in the
126: %concentration profile of Pt as obtained by
127: %Auger electron spectroscopy (AES) depth profiling analysis.
128:
129:
130:
131:
132:
133:
134:
135:
136:
137: \section{The experimental setup}
138:
139: Although the experimental setup has been outlined in ref. \cite{Sule_condmat06b},
140: however, we briefly also give the most important features of the
141: measuremnts.
142: The experimental setup is the following:
143: According to the crossectional TEM (XTEM) results the thickness of the layers in
144: samples are:
145: Pt $13$ nm/Ti $11$
146: nm/Si substrate
147: (denoted throughout the paper as Pt/Ti), and Ta 21 nm (cap layer to prevent oxidation
148: of Ti)/Ti 11 nm/Pt 12 nm/Si substrate (denoted as Ti/Pt).
149: The XTEM images are shown in Fig 1.
150: %are sharp and straight and no asymmetry is observable.
151: For the sake of simplicity we consider our multilayer samples as bilayers and we study the
152: atomic transport processes at the Ti/Pt and Pt/Ti interfaces.
153: Both samples have been AES depth profiled by applying various sputtering conditions.
154: The sample has been rotated during sputtering. In the following we will outline results of $500$ eV
155: Ar$^+$ ion bombardment at an
156: angle of incidence of $10^{\circ}$ (with respect to the surface).
157: The atomic concentrations of Pt, Ti, Ta and Si were calculated by the relative sensitivity method
158: taking the pure material's values from the spectra. The oxygen atomic concentration has been
159: calculated by normalizing the measured oxygen Auger peak-to-peak amplitude to TiO$_2$
160: \cite{Vergara}. The depth scale was determined by assuming that the sputtering yield ($Y_i$) in the
161: mixed layer is the weighted sum of the elemental sputtering yields ($\sum_i X_i Y_i$).
162: %Two typical AES depth profiles recorded on samples of Pt/Ti and Ti/Pt at $500$ eV ion energy,
163: % are shown, respectively in Fig. 2. It is clear without any detailed evaluation that the Pt/Ti
164: %and Ti/Pt transitions are very different.
165: %While the Ti/Pt transition
166: %is a "normal" one, in case of Pt/Ti an unusual deep penetration of Pt to the Ti phase is observed.
167:
168: % We have carried out several experiments, using different parts of the sample
169: %and applying various sputtering conditions.
170: %In all cases we recognized oxygen. The oxygen concentration varied considerably.
171: %The oxygen at the metal/substrate interface is due to the native
172: %oxide on the Si substrate.
173: %Since the bulk level of oxygen slightly correlated with the ion current intensity
174: %(the larger
175: %the ion current intensity the lower the oxygen AES signal) part of the oxygen is
176: %contamination occurring during the AES depth profiling process.
177: %On the other hand the interface broadening does not
178: %correlate with the
179: %concentration of
180: %the oxygen
181: %in the Ti/Pt interface.
182: %Thus we conclude
183: %that the atomic transport is not affected by the presence of the slight oxygen contaminant.
184:
185: % XTEM give reasonable good estimate of the roughness of the as received sample, which was found to be about $\sim 1,5$ nm (intermixing would result in a gradual change of the contrast; since it was missing we estimate that the initial intermixing is less than $\sim 1,5$ nm). Ion-sputtering might cause surface roughening as well as ion mixing. In the case of our experimental conditions (rotated sample, grazing angle of incidence, low relative sputtering yield) the ion bombardment induced roughening is expected to be weak, but cannot be ruled out; thus part of the measured total broadening might be due to roughening.
186: The broadening of the interface is frequently characterizied by the depth resolution. The depth resolution is defined as the distance of points on the depth profile exhibiting 84 \% and 16 \% concentrations. This definiton has been introduced for the "normal" cases, when either the ion mixing or the roughening can be described by a Gaussian convolution resulting
187: in an erf fuction transition in the depth profile. The same definition used for other cases as well, however.
188:
189: If the transition does differ from the $erf$ function (e.g. when the mobility of one of the components of a diffusion couple is much higher than that of the other's) one might give also the distances between points of $84$ \% and $50$ \%, and $50$ \% and $16$ \%.
190: The ratio of these distances gives us the asymmerty of intermixing (shown in Table 1).
191:
192:
193:
194: \section{The setup of the atomistic simulations}
195:
196: In order to get more insight to the mechanism of interdiffusion
197: classical molecular dynamics simulations have also been used to simulate the ion-solid interaction
198: (using the PARCAS code \cite{Nordlund_ref}).
199: Here we only shortly summarize the most important aspects.
200: A variable timestep
201: and the Berendsen temperature control is used to maintain the thermal equilibrium of the entire
202: system. \cite{Allen}. The bottom layers
203: are held fixed in order to avoid the rotation of the cell.
204: Periodic boundary conditions are imposed laterarily and a free surface is left for the ion-impacts.
205: The temperature of the atoms in the
206: outermost layers was softly scaled towards the desired temperature to provide temperature control and ensure
207: that the pressure waves emanating from cascades were damped at the borders.
208: The lateral sides of the cell are used as heat sink (heat bath) to maintain the thermal equilibrium of the entire
209: system \cite{Allen}.
210: The detailed description of other technical aspects of the MD simulations are given in \cite{Nordlund_ref,Allen} and details specific to the current system in recent
211: communications \cite{Sule_NIMB04,Sule_PRB05,Sule_SUCI,Sule_NIMB04_2}.
212:
213: We irradiate the bilayers Pt/Ti and Ti/Pt
214: with 0.5 keV Ar$^+$ ions repeatedly with a time interval of 10-20 ps between each of
215: the ion-impacts at 300 K
216: which we find
217: sufficiently long time for the termination of interdiffusion, such
218: as sputtering induced intermixing (ion-beam mixing) \cite{Sule_NIMB04}.
219: The initial velocity direction of the
220: impacting ion was $10$ degrees with respect to the surface of the crystal (grazing angle of incidence)
221: to avoid channeling directions and to simulate the conditions applied during ion-sputtering.
222: We randomly varied the impact position and the azimuth angle $\phi$ (the direction of the ion-beam).
223: In order to approach the real sputtering limit a large number of ion irradiation are
224: employed using automatized simulations conducted subsequently together with analyzing
225: the history files (movie files) in each irradiation steps.
226: In this article we present results up to 200 ion irradiation which we find suitable for
227: comparing with low to medium fluence experiments. 200 ions are randomly distributed
228: over a $20 \times 20$ \hbox{\AA}$^2$ area which corresponds to $\sim 5 \times 10^{15}$
229: ion/cm$^2$ ion fluence
230: and the removal of few MLs.
231: % Computer atomistic simulations have also been used recently for simulating
232: %low-energy ion-sputtering \cite{Karolewski,Thijsse,Zhong,Hanson}.
233:
234:
235: The size of the simulation cell is $110 \times 110 \times 90$ $\hbox{\AA}^3$ including
236: 57000 atoms (with 9 monolayers (ML) film/substrate).
237: At the interface (111) of the fcc crystal is parallel to (0001) of the hcp
238: crystal
239: and the close packed directions are parallel.
240: The interfacial system is a heterophase bicrystal and a composite object of
241: two different crystals with different
242: symmetry is created as follows:
243: the hcp Ti is put by hand on the (111) Pt bulk (and vice versa) and various structures are probed
244: and are put together randomly. Finally the one which has the smallest
245: misfit strain prior to the relaxation run is selected.
246: The difference between the width of the overlayer and the bulk does not exceed $2
247: -3$ \hbox{\AA}.
248: The remaining misfit is properly minimized below $\sim 6 \%$ during the relaxation
249: process so that the Ti and Pt layers keep their original crystal structure and we
250: get an
251: atomically sharp interface.
252: During the relaxation (equilibration) process the temperature is softly scaled down
253: to zero.
254: According to our practice we find that during the temperature scaling down the
255: structure
256: becomes sufficiently relaxed therefore no further check of the structure has been
257: done.
258: Then the careful heating up
259: of the system to $300$ K has been carried out. The systems were free from any serious built-in strain
260: and the lattice mismatch is minimized to the lowest possible level.
261: The film and the substrate are $\sim 20$ and $\sim 68$ $\hbox{\AA}$ thick, respectively.
262:
263: In order to reach the most efficient ion energy deposition at the interface,
264: we also initialize recoils placing the ion above the interface by $10$ \hbox{\AA} (and
265: below the free surface in the 9 ML thick film) also at grazing angle of incidence
266: ($10^{\circ}$ to the surface)
267: with $500$ eV ion energy.
268: In this way
269: we can concentrate directly on the intermixing phenomenon avoiding
270: many other processes occur at the surface (surface roughening, sputter erosion, ion-induced surface diffusion, cluster ejection, etc.) which weaken energy deposition at the interface.
271: Further simplification is that channeling recoils are left to leave the cell
272: and in the next step these energetic and sputtered particles are deleted.
273:
274: We used a tight-binding many body potential, developed by Cleri and
275: Rosato (CR) on the basis of the second moment approximation to the density of states \cite{CR}, to describe interatomic interactions.
276: This tight-binding approach is formally analogous to the embedded-atom method (EAM) \cite{EAM}.
277: This type of a potential gives a good description of lattice vacancies, including atomic migration
278: properties and a reasonable description of solid surfaces and melting \cite{CR}.
279: Since the present work is mostly associated with the elastic properties,
280: melting behaviors, interface and migration energies, we believe the model used should be suitable for this study.
281: The interatomic interactions are calculated up to the 2nd nearest neighbors
282: and a cutoff is imposed out of this
283: region.
284: This amounts to the maximum interatomic distance of $\sim 6$ \hbox{\AA}.
285: For the crosspotential of Ti and Pt we employ an interpolation scheme \cite{Sule_PRB05,Sule_SUCI,ZBL}
286: between the respective elements.
287: The CR elemental potentials and the interpolation scheme for heteronuclear interactions
288: have widely been used for MD simulations \cite{Sule_PRB05,Stepanyuk2,Goyhenex,Levanov}.
289: The Ti-Pt interatomic crosspotential of the Cleri-Rosato \cite{CR} type is fitted to the experimental
290: heat of
291: mixing of the corresponding alloy system \cite{Sule_NIMB04,Sule_NIMB04_2}.
292: The scaling factor $r_0$ (the heteronuclear first neighbor distance) is calculated as the average of the elemental first neighbor distances.
293:
294:
295:
296: The computer animations can be seen in our web page \cite{web}.
297: Further details are given in \cite{Nordlund_ref} and details specific to the current system in recent
298: communications \cite{Sule_PRB05,Sule_NIMB04}.
299:
300:
301:
302: \section{Results}
303:
304:
305:
306: In Fig 3 the evolution of the sum of the square of atomic displacements (SD) of all intermixing atoms $\langle R^2 \rangle= \sum_i^N [{\bf r_i}(t)-{\bf r_i}(t=0)]^2$, obtained by molecular dynamics simulations, where (${\bf r_i}(t)$ is the position vector of atom 'i' at time $t$, $N$ is the total number of atoms included in the sum), can be followed as a function of the ion fluence.
307: Lateral components ($x,y$) are excluded from
308: $\langle R^2 \rangle$ and only contributions from intermixing atomic displacements perpendicular to the layers
309: are included ($z$ components). We follow during simulations the time evolution
310: of $\langle R^2 \rangle$ which reflects the atomic migration through the interface (no other
311: atomic transport processes are included).
312: Note, that we do not calulate the mean square of atomic displacements (MSD) which is an
313: averaged SD over the number of atoms included in the sum (MSD=$\langle R^2 \rangle/N$).
314: MSD does not reflect the real physics when localized events take place, e.g. when only few dozens of atoms
315: are displaced and intermixed.
316: In such cases the divison by $N$, when $N$ is the total number of the atoms in the simulation cell
317: leads to the meaningless $\langle R^2 \rangle/N \rightarrow 0$ result when $N \rightarrow \infty$,
318: e.g. with the increasing number of atoms in the simulation cell.
319: Also, it is hard to give the number of "active" particles which
320: really take place in the transient atomic
321: transport processes.
322: %------------------------------------------------------
323: \begin{figure}[hbtp]
324: \begin{center}
325: \includegraphics*[height=6cm,width=6.9cm,angle=0.]{fig2a.eps}
326: \includegraphics*[height=6cm,width=6.9cm,angle=0.]{fig2b.eps}
327: \caption[]{
328: The concentration depth profile as a function of the removed layer thickness (nm)
329: obtained by AES depth profiling analysis
330: using ion-sputtering at 500 eV Ar$^+$ ion energy ($10^{\circ}$ grazing angle of incidence with respect to the surface) in P
331: t/Ti (upper Fig: 1a) and Ti/Pt (lower Fig: 1b).
332: }
333: \label{fig2}
334: \end{center}
335: \end{figure}
336: %------------------------------------------------------
337: Hence we prefer to use the more appropriate quantity SD.
338: In Fig 3 we present
339: $\langle R^2 \rangle$ as a function of the number of ion impacts $N_i$ (ion number fluence).
340: $\langle R^2 \rangle (N_i)$ corresponds to the final value of
341: $\langle R^2 \rangle$
342: obtained during the $N_i$th simulation. The final relaxed structure of the simulation of the
343: $(N_i-1)$th ion-bombardment is used as the input structure for the $N_i$th ion-irradiation.
344: The
345: asymmetry of
346: mixing can clearly be seen when $\langle R^2 \rangle (N_i)$
347: and the depth profiles are compared in
348: Ti/Pt and in Pt/Ti in Figs 1a,1b and 2.
349: The computer animations of the simulations together with the plotted broadening values
350: at the interface in inset
351: Fig 3 of ref. \cite{Sule_condmat06b} also reveal the stronger
352: interdiffusion in Pt/Ti \cite{web}.
353:
354:
355:
356: As it has already been shown in ref. \cite{Sule_condmat06b}
357: a relatively weak intermixing is found in Ti/Pt ($\sigma \approx 20$ $\hbox{\AA}$) while an unusually high
358: interdiffusion occurs in the Pt/Ti bilayer ($\sigma \approx 70$ $\hbox{\AA}$).
359: MD simulations provide $\sim 8$ ML ($\sim 20$ $\hbox{\AA})$ and $\sim 16$ ML ($\sim 40$ $\hbox{\AA})$ thick interface after $200$ ion impacts, respectively.
360: The comparison of the measured $\sigma$ with the simulated broadening
361: using the $84-16$ \% rule in both cases can only be carried out with great care
362: \cite{Sule_condmat06b}.
363: %In principle, these values are not comparable directly.
364: %However, we make some simple assumptions at this point.
365: %We expect that the interface roughening has smaller contribution to broadening
366: %than intermixing.
367: %This assumption holds because if the wavelenght of interface waving would be higher than the nanoscale
368: %than the measured $\sigma$ should be larger than few tens of $\hbox{\AA}$.
369: %If the wavelenght and the amplitude of interface waving is in the nanoscale than
370: %$\sigma$ could come mostly from roughening.
371: %Unfortunately we have no results for the roughening of the samples after ion-bombardment.
372: %However, we expect that if the simulated $\sigma_{mix}$ is comparable with the measured
373: %$\sigma$, than we can expect that the ion-sputtered $\sigma_{ro} \ll \sigma_{mix}$.
374:
375: Moreover, the applied setup of the simulation cell, in particular the $20$ $\hbox{\AA}$ film thickness is assumed to be
376: appropriate for simulating broadening.
377: Our experience shows that the variation of the film thickness does not affect the
378: final result significantly, except if ultrathin film is used (e.g. if less than
379: $\sim 10$ $\hbox{\AA}$ thick film). At around $5$ or less ML thick film surface roughening could affect
380: mixing, and vice versa \cite{Sule_NIMB04_2}.
381: Also, we do not carry out complete layer-by-layer removal as in the experiment.
382: It turned out during the simulations that the ions mix the interface the most efficien
383: tly
384: when initialized $\sim 10 \pm 3$ $\hbox{\AA}$ above the interface.
385: This value is naturally in the range of the projected range of the ions.
386: Hence, the most of the broadening is coming from this regime of ion-interface distance
387: .
388: Initializing ions from the surface it takes longer and more ions are needed to
389: obtain the same level of damage and broadening at the interface as it has been found b
390: y ion-bombarding
391: from inside the film.
392: %We demonstrate the comparativity of the simulated and measured broadenings for the Cu/
393: %Co bilayer
394: %elsewhere using this way of simulation technique of ion-sputtering
395: %\cite{CuCo}.
396:
397: In ref. \cite{Sule_condmat06b} we have shown that
398: the measured ion-sputtering induced broadening of $\sigma
399: \approx 20$ $\hbox{\AA}$ for Ti/Pt is in nice agreement with
400: the simulated value.
401: Hence we expect that the most of the measured $\sigma$ is coming from intermixing
402: and interface roughening contributes to $\sigma$ only slightly.
403: The nice agreement could be due to that the saturation of intermixing (broadening) dur
404: ing ion-sputtering is insensitive to
405: the rate of mixing in the as received samples
406: ($\sigma_0 \approx 15$ $\hbox{\AA}$ interface width (including the interface roughenin
407: g) in both samples).
408: This is because
409: a sharp interface and a weakly mixed one (before bombardment in the as-received samples) lead to the same magnitude of broadening
410: upon ion-sputtering, because the binary systems reach the same steady state
411: of saturation under the same conditions (ion energy, impact angele, etc.).
412: This is rationalized by our finding that
413: during simulations we start from a sharp interface ($\sigma_0 \approx 0$) and
414: we get a very similar magnitude of $\sigma$ than by AES.
415:
416: % Asymmetric AES depth profiles (when the broadening of A/B interface is different from that of B/A) have already been observed \cite{Hofmann,Barna}.
417: %The asymmetric behavior has been explained by the large relative sputtering yield of the elements (preferential sputtering)
418: %\cite{Hofmann,Barna,Berg}.
419: %The asymmetry of intermixing is also known during thin film growth
420: %\cite{Buchanan,MM}.
421: %In the present experiment the relative sputtering yield of
422: %$Y_{Pt}/Y_{Ti} \approx 0.7$ at $500$ eV (at $1500$ eV we find $Y_{Pt}/Y_{Ti} \approx 0.9$ and a similar rate of broadening)
423: %therefore the mechanism is different.
424: %This kind of an ion-bombardment induced asymmetric mixing in metallic diffusion couples has not reported yet
425: %at best of our knowledge.
426:
427:
428: % According to the simulations, the mixing of the Pt increases with the fluence,
429: %at $200$ ions irradiation (which is the highest we reached because of the
430: %limited CPU time) $\sigma \approx 40$ $\hbox{\AA}$ was found. This is
431: %not a saturation value (in contrast to the Ti/Pt case) hence
432: %we can expect further increase in $\sigma$.
433: %The simulated values of broadening are purely coming from intermixing.
434: %We expect that the rest of measured $\sigma$ is coming from roughening.
435: %Nevertheless, MD simulations reproduce the mixing asymmetry and
436: %we are able to capture the essential features of the phenomenon.
437:
438: \section{Discussion}
439:
440: \subsection{The nonlinear scaling of $\langle R^2 \rangle$}
441:
442: In recent papers we explained single-ion impact induced intermixing as an interdiffusion
443: process governed by the mass-anisotropy parameter (mass ratio) in these bilayers
444: \cite{Sule_PRB05,Sule_SUCI}.
445: We also present $\langle R^2 \rangle$ results for Cu/Co bilayer in Fig 2.
446: Ti/Pt and Pt/Ti are mass-anisotropic, while Cu/Co is
447: a mass-isotropic bilayer.
448: Cu/Co can be used as a reference system, since it is composed of immiscible
449: elements, hence a weak rate of intermixing is expected for.
450: It has already been shown in refs. \cite{Sule_NIMB04,Sule_PRB05} that the backscattering of the hyperthermal particles (BHP)
451: at the heavy interface leads to the increase in the energy density of the collisional cascade.
452: %It should also be noted that mass-effect induced preferential sputtering phenomena, such as
453: %sputtering yield amplification in metal bilayers have already been reported \cite{Berg}.
454: %However, these studies did not seem to focus on the effect of mass-anisotropy on intermixing.
455: %------------------------------------------------------
456: \begin{figure}[hbtp]
457: \begin{center}
458: \includegraphics*[height=6cm,width=7.5cm,angle=0.]{fig2.eps}
459: \caption[]{
460: The simulated square of intermixing atomic displacements $\langle R^2 \rangle$ ($\hbox{\AA}^2$) in Pt/Ti, Ti/Pt and in Cu/Co as a function of the
461: ion-fluence (number of ions) obtained during the ion-sputtering of these bilayers at 500 eV ion energy (results are shown up to 100 ions).
462: The dotted lines (iso-Pt/Ti and iso-Ti/Pt) denote the results obtained for the artificial mass-isotropic
463: Pt/Ti and Ti/Pt bilayers, respectively.
464: The ions are initiated from the surface.
465: {\em Inset}:
466: The simulated broadening at the interface in $\hbox{\AA}$ as a function of
467: the number of ions at 500 eV ion energy.
468: The ions are initiated from $10$ $\hbox{\AA}$ above the interface.
469: The error bars denote a statistical uncertainty in the measure of broadening.
470: }
471: \label{fig3}
472: \end{center}
473: \end{figure}
474: %------------------------------------------------------
475: We have found that the jumping rate of atoms through the interface is seriously affected by the
476: mass-anisotropy of the interface when energetic atoms (hyperthermal particles) are present and which leads to the
477: preferential intermixing of Pt to Ti \cite{Sule_NIMB04,Sule_PRB05}.
478:
479:
480:
481: According to the BHP model,
482: in Pt/Ti, the energetic Ti light particles are backscattered downwards and confined
483: below the heavy interface
484: in the bulk which results in increasing energy density below the interface.
485: In Ti/Pt we find the reversed case: the Ti atoms are backscattered upwards, towards the
486: nearby free surface (the ion-sputtered surface is always close to the interface when the
487: interface is irradiated).
488: The injection of the Pt particles to the Ti phase can be seen in lower Fig. 4.
489: %Surprisingly we find no serious difference in the average lifetime of the collisional cascade and thermal spike in Pt/Ti
490: %and in Ti/Pt.
491: We know from ref. \cite{Sule_NIMB04} that in Ti/Pt thermal spike occurs also at low ion energies.
492: At $0.5$ keV energy we also find thermal spike with the average lifetime of $\tau \approx 5$ ps averaged
493: during few tens of consequtive events.
494: This is due to the backscattering of light particles at the interface as it has been shown
495: in refs. \cite{Sule_PRB05,Sule_NIMB04}.
496: Hence mass anisotropy enhances thermal spike in both bilayers.
497: In Pt/Ti, however we find the further enhancement of the heat spike and
498: $\tau \approx 10$ ps.
499: Unfortunatelly we cannot explain further enhancement of interdiffusion
500: in Pt/Ti by simple mass effect.
501: We reach the conclusion that there must be another accelerative field which
502: speeds up Pt particles to ballistic transport.
503: This is reflected by the divergence of $\langle R^2 \rangle$ from linear scaling in Fig 3. for
504: Pt/Ti.
505: Interdiffusion takes place via ballistic jumps (ballistic mixing), when $\langle R^2 \rangle$ grows
506: asymptotically as $N^2$, where $N$ is the number fluence
507: (the same asymptotics holds as a function ion-dose or ion-fluence).
508: This can clearly be seen in Fig. 3 for Pt/Ti.
509: The horisontal axis is proportional to the time of ion-sputtering, hence
510: $\langle R^2 \rangle
511: \propto t^2$ which is the time scaling of ballistic atomic transport
512: \cite{anomalous}.
513: In our particular case $1$ ion-bombardment corresponds
514: to $t \sim 10$ ps which we find sufficient time for the evolution of $\langle R^2 \rangle$.
515: Anyhow, above this $t$ value the asymptotics of $\langle R^2 \rangle (t)$ is invariant
516: to the choice of the elapsed time/ion-bombardment induced evolution of $\langle R^2 \rangle (t)$.
517: Hence the transformation between ion-fluence and time scale is allowed.
518: $\langle R^2 \rangle \propto t^2$ and $\langle R^2 \rangle \propto t$
519: time scalings have been found even for the single-ion impacts averaged for few events (when $\langle R^2 \rangle (t)$ is plotted only for a single-ion impact) for Pt/Ti and Ti/Pt, respectively.
520:
521: No such ballistic behavior can be seen for Ti/Pt in Fig 3.
522: In Ti/Pt we find $\langle R^2 \rangle \propto t$ time scaling which is
523: due to the shorter average lifetime of the collisional cascades \cite{Sule_PRB05,Sule_NIMB04}.
524: The mean free path of the energetic particles are much shorter in Ti/Pt (this can be seen
525: qualitatively in Fig 4. if we compare the length of the atomic trajectories for Pt between
526: upper and lower panels of Fig 4.).
527: The $\langle R^2 \rangle \propto t^2$ scaling used to be
528: considered as the signature of
529: anomalous diffusion (superdiffusion) in the literature \cite{anomalous}.
530: In the upper panel of Fig 2. the fingerprint of superdiffusive feature
531: of intermixing is detected by AES as a tail in the concentration profile of Pt.
532: No such tail occurs for Ti/Pt in the lower panel of Fig 2. where the profile
533: can be characterized by "normal" $erf$ functions.
534: Superdiffusive features, however, have never been reported before for intermixing, only
535: for e.g. random walk of adatoms (Levy flight) on solid surfaces \cite{anomalous}.
536: Therefore, the Pt/Ti system is highly unusual and there has to be more surprise in store.
537: The anomalous impurity diffusion of N in stainless steel \cite{Abrasonis} and
538: the observed large interdiffusion depths in various transition metal/Al diffusion
539: couples \cite{Buchanan} could also be understood as super-interdiffusive processes.
540: %------------------------------------------------------
541: \begin{figure}[hbtp]
542: \begin{center}
543: \includegraphics*[height=5.5cm,width=6.9cm,angle=0.]{fig4a.eps}
544: \includegraphics*[height=5.5cm,width=6.9cm,angle=0.]{fig4b.eps}
545: \caption[]{
546: The crossectional view of a typical collisional cascade at the interface
547: with atomic trajectories (two monolayers are shown at the interfaces as a crossectional
548: slab cut in the middle of the simulation cell)
549: in Ti/Pt (upper panel) and in Pt/Ti (lower panel).
550: The positions of the energetic particles are collected up to 500 fs during
551: a $500$ eV single ion-impact event initialized $10$ $\hbox{\AA}$ above the interface
552: using the same conditions as in other simulations.
553: The vertical axis corresponds to the depth position given in $\hbox{\AA}$.
554: The position $z=0$ is the depth position of the interface.
555: }
556: \label{fig4}
557: \end{center}
558: \end{figure}
559: %------------------------------------------------------
560: The trajectories of the reversed Ti recoils can be seen in upper Fig. 4 and
561: no intermixing Ti atomic positions can be found in upper panel of Fig 4 (Ti/Pt).
562: Although, Fig. 4 has no any statistical meaning (as it should be), however,
563: the trajectories are selected among typical events hence some useful information
564: can be obtained for the transport properties of energetic Pt atoms.
565: In lower panel of Fig 4. we can see the ballistic trajectories of intermixing hyperthermal Pt atoms (Pt/Ti).
566: The reversed fluxes of the Ti particles at the interface
567: and the weaker intermixing of Pt atoms to the Ti phase
568: result in
569: the weaker intermixing in Ti/Pt than
570: in the Pt/Ti system.
571: This is nicely reflected in the $\langle R^2 \rangle$ of Ti/Pt which remains in the range of Cu/Co.
572: Hence Fig. 4 depicts us at atomistic level what we see in the more statistical quantity
573: $\langle R^2 \rangle$. In Ti/Pt we see much shorter inter-layer atomic trajectories
574: while in Pt/Ti ballistic trajectories of Pt atoms can be seen (going through the interface).
575:
576: %------------------------------------------------------
577: \begin{figure}[hbtp]
578: \begin{center}
579: \includegraphics*[height=6cm,width=6.9cm,angle=0.]{fig5.eps}
580: \caption[]{
581: The Cleri-Rosato potential energy (eV) as a function of the
582: interatomic distance ($\hbox{\AA}$) in Ti and in Pt averaged for
583: an atom with its 12 first neighbors in the hcp and fcc lattices, respectively.
584: }
585: \label{fig5}
586: \end{center}
587: \end{figure}
588: %------------------------------------------------------
589:
590:
591: \subsection{The effect of mass and size anisotropy}
592:
593:
594:
595: In order to clarify the mechanism of intermixing and to understand how much
596: the interfacial anisotropy influences IM
597: simulations have been carried out with mass ratio $\delta$ is artificially set to $\delta
598: \approx 1$ (mass-isotropic, ions are initiated from the surface). We find that $\langle
599: R^2 \rangle$ is below the corresponding curve of Pt/Ti (see
600: Fig 3, iso-Pt/Ti, dotted line, and animation \cite{web}).
601: The $\langle R^2 \rangle$ scales nearly linearily as a function of the number of ions
602: (and with $t$) for iso-Pt/Ti.
603: The asymptotics of $\langle R^2 \rangle$ is sensitive to the mass ratio.
604: Hence we reach the conclusion that the mass-effect is robust,
605: and the energetic particle scattering mechanism is weakened
606: when
607: mass-anisotropy
608: vanishes and the magnitude of intermixing is also weakened.
609: %This finding together with our AES measurements (with the long-range tail) confirms our recent results reported for
610: %various bilayers in which a strong correlation has been obtained between
611: %the experimental and simulated mixing efficiencies and mass anisotropy in various metallic
612: %bilayers
613: %\cite{Sule_PRB05}.
614: %In this article we found that below a certain treshold mass ratio value ($\delta \le 0.33$)
615: %the rate of intermixing increases abruptly \cite{Sule_PRB05}.
616: %On the basis of the results obtained in this paper this
617: %surprising interdiffusive behaviour of bilayers
618: %could be explained by the anomalous nature of mixing which can be tuned by
619: %the mass-anisotropy (mass ratio) of the systems.
620: %The experimentally observed mixing asymmetry can also partly be explained by
621: %mass effect.
622:
623:
624: The same type of MD simulations has been carried out in Cu/Co
625: as well, and the results are shown in Fig 3. The Cu/Co system is mass-isotropic,
626: and thus we expect that the magnitude of $\langle R^2 \rangle$ obtained for iso-Ti/Pt and iso-Pt/Ti should be
627: similar to that of Cu/Co.
628: It turns out, however, that the $\langle R^2 \rangle$
629: of iso-Ti/Pt and Ti/Pt are higher than that of Cu/Co. The $\langle R^2 \rangle$
630: of iso-Ti/Pt is somewhat even higher than that of Ti/Pt.
631: Hence the effect of mass-isotropy on $\langle R^2 \rangle$ is opposite in Ti/Pt
632: than in Pt/Ti.
633: This might be due to the reversed fluxes of Ti recoils in the mass-anisotropic Ti/Pt,
634: shown in upper panel of Fig 4.
635:
636:
637: To understand the physical origin of the higher intermixing in iso-Pt/Ti
638: than in Cu/Co,
639: we should consider a more complex anisotropy of the interface
640: then previously
641: expected.
642: We can rule out the effect of thermochemistry (e.g. heat of mixing $\Delta H$)
643: because
644: it has also been shown recently, that $\Delta H$ has no
645: apparent effect on intermixing during single-ion impact in this bilayer \cite{Sule_NIMB04}.
646: Also, thermochemistry does not explain the mixing asymmetry observed in Ti/Pt and
647: in Pt/Ti since $\Delta H$ must be the same in the two systems.
648: Moreover, simulated ion-sputtering with $\Delta H \gg 0$ (switching off the attractive term of the cross-potential
649: and using a repulsive (Born-Mayer) Ti-Pt interaction potential for Pt/Ti)
650: leads to intermixing instead of the expected sharp interface (see also animations \cite{web}).
651: We reach the conclusion that there must be a strong accelerative effect in the system which even
652: counterbalance the effect of heteronuclear repulsion.
653:
654: The employed interaction potential energy functions are plotted
655: in Fig ~(\ref{fig5}) as a function of the interatomic distance for Ti and for Pt. It can clearly be seen that
656: the thermal atomic sizes (the width of the potential valley at higher atomic kinetic energy) are different.
657: This difference could not be realized during the short time period of a collisional cascade
658: ($\tau < 0.5$ ps).
659: However, the increased lifetime of the thermal spike in these bilayers allows few vibrational periods
660: and which is sufficient for the onset of the effect of atomic size anisotropy.
661: The atomic radius (the width of the vibration around the equilibrium lattice site) of the Ti atoms
662: is larger than that of in Pt. The atomic size anisotropy is even more robust at 1-2 eV energies above the energy minimum (hyperthermal atomic energies, see the horizontal lines in Fig 4). This explains the fact that the smaller and heavier Pt atoms are the first
663: ballistic diffuser during ion-beam mixing \cite{Sule_NIMB04}.
664: The atomic size difference and the mass-anisotropy together could explain at least partly the stronger simulated (and measured) intermixing in Pt/Ti and in Ti/Pt than
665: in Cu/Co. In the latter mass-isotropic bilayer no serious atomic size difference (anisotropy) is found (see the inset Fig ~(\ref{fig5})).
666: The mixing in the iso-Pt/Ti could also be understood by the size anisotropy:
667: the larger energetic Ti atoms are reversed at the interface and confined in the bulk.
668: The smaller energetic Pt atoms, however, are injected to the Ti bulk during
669: the cascade period.
670:
671:
672:
673:
674: \subsection{Nonlinear accelerative field}
675:
676:
677: Since it is evident from the analysis of the previous subsections that mass and size
678: anisotropy do not fully acount for the enhancement of IM in Pt/Ti, there must be
679: another source(s) of the divergence of $\langle R^2 \rangle$.
680: The nearly parabolic time scaling of $\langle R^2 \rangle$ suggests ballistic
681: IM induced by intrinsic acceleration of IM Pt particles.
682: The following simple picture seems to explain the nonlinear time scaling
683: of $\langle R^2 \rangle$:
684: Instead of the rapid solidification of the thermal spike intrinsic lattice
685: effects speed up inter-layer transport of Pt atoms until
686: saturation occurs in the Ti phase.
687:
688:
689:
690: Normal (thermally activated) atomic transport processes have already been carefully ruled out
691: in ref. \cite{Abrasonis} for the anomlaous diffusion of N in nitrided austenitic stainless steel.
692: There is no specific reason to expect that these reasonings do not hold for Pt/Ti.
693: Namely, we rule out that vacancy, multivacancy or self-interstitial mediated, stress induced or concentration
694: dependent, thermal acceleration diffusion could explain the occurence of the asymmetry and the anomalous
695: nature of interdiffusion in Pt/Ti.
696:
697:
698: Although, the anisotropy present in the bilayer systems have serious effect on intermixing,
699: the asymmetry of IM still remains unexplained.
700: This seems to be puzzling at first sight, however, we do believe that similar nonlinear effects
701: are responsible for the large diffusion depth of Pt as it has been proposed by
702: Abrasonis {\em et al.} in ref. \cite{Abrasonis}.
703: Similar nonlinear forces induced accelerative mechanism has been found the most recently for other diffusion couples
704: \cite{Sule_condmat}.
705: On the basis of these findings
706: anharmonic lattice vibrations (intrinsic localized modes) could promote atomic mobility.
707: It has already been shown that the localization of vibrational energy within an extended lattice
708: can influence various materials properties \cite{Swanson}.
709:
710: Since the enchanced atomic mobility is also present in weakly repulsive heteronuclear field
711: (with repulsive Ti-Pt internuclear potential)
712: one might expect that the accelerative field is rather strong and could indeed be due to
713: the intrinsic localization of anharmonic vibration modes at the interface region.
714: The fundamental question remains to be answered yet that
715: if nonlinear and long-range forces drive superdiffusion,
716: why these nonlocal anharmonic vibrational modes are active in Pt/Ti and why they are suppressed in Ti/Pt.
717: Using repulsive Ti-Pt potential in Ti/Pt we find the further
718: suppression of IM hence no accelerative long-range forces are active in this
719: material which could not surpass the slow down of the particles in the
720: repulsive interparticle field.
721:
722:
723: The asymmetry of accelerative effects has also been found in Pt/Al couple during simulated
724: vapor deposition of Pt on Al(111). No transient intermixing is found for Al deposition on Pt(111) \cite{Sule_condmat}.
725: In this case we explained the apparent asymmetry by atomic size mismatch \cite{Sule_condmat}.
726: The ultrafast diffuser specie must be the smaller one. Moreover the host material atoms must be not only
727: larger but also anharmonic. This is indeed also the case for Ti.
728: Hence anharmonic vibrational modes could be excited in the
729: anharmonic host material which exhibits accelerative field.
730: In the reverse case the larger particles could not excite the nearly harmonic host material with smaller particles.
731:
732: It should also be noted that the intrinsic accelerative effects of the anharmonic lattice modes
733: accelerate moving particles during the cascade and thermal spike period increasing the lifetime of
734: the heat spike. Few of these accelerated particles could reach hyperthermal kinetic energies (hot atoms).
735: According to our simulations in Pt/Ti
736: the mean free path of these particles is in the range of $\sim 10$ $\hbox{\AA}$.
737: More careful analysis certainly will be necessary, however, to identify the details of this
738: specific mechanism which could also hold for couple of other diffusion couples.
739:
740:
741: \section{Conclusions}
742:
743: We could reproduce by means of MD simulations the experimentally observed mixing asymmetry between Ti/Pt and Pt/Ti.
744: We find a robust mass effect on interfacial mixing in Pt/Ti, although
745: the mass effect does not fully account for intermixing in Pt/Ti:
746: the atomic size and mass-anisotropy are together could be responsible for
747: the observed strong interfacial mixing in Pt/Ti and
748: the weaker mixing in Cu/Co is explained by atomic size and mass-isotropy.
749: We get a nice agreement for interface broadening in Ti/Pt with experiment while for
750: Pt/Ti the discrepancy is not negligible.
751: We conclude from this that interface roughening might has a significant
752: contribution to broadening in Pt/Ti.
753:
754: We also find that the sum of the squares of atomic displacements ($\langle R^2 \rangle$) through the anisotropic interface
755: scales nonlinearily in Pt/Ti ($\langle R^2 \rangle \propto t^2$) as a function of the time (and the ion-number fluence)
756: as shown in Fig. 3.
757: From the upper panel of Fig. 2 we see the experimental fingerprint of ballistic mixing (super-interdiffusion)
758: as a well developed tail in the concentration profile of Pt with an approximate length of $\sim 25$
759: $\hbox{\AA}$ along the depth direction.
760: The long range (high diffusity) tail in the AES spectrum provides a compelling evidence for the superdiffusive transport
761: process of Pt atoms.
762: In Ti/Pt a nearly linear scaling ($\langle R^2 \rangle \propto t$) is found which is due to the shorter average
763: lifetime of the collisional cascade period and which leads to
764: shorter mean free path of the energetic Pt particles.
765: The lack of
766: a tail in the AES concentration profile for Ti/Pt (lower panel of Fig 2.) indicates a
767: concentional (thermally activated) mechanism of interdiffusion in Ti/Pt.
768: In this system ion-sputtering increases broadening only slightly: in the as received sample
769: we find $\sigma_0 \approx 15$
770: $\hbox{\AA}$
771: , and $\sigma \approx 20$ $\hbox{\AA}$ is measured after AES depth profiling.
772: Based on atomistic simulations we find that
773: the backscattering of the sizeable and light Ti particles at the interface suppresses further
774: intermixing.
775:
776: We conclude that the observed and simulated long range depth distribution of Pt atoms
777: in the Ti phase of Pt/Ti cannot be understood by any established mechanisms of radiation-enhanced
778: diffusion.
779: In Pt/Ti, we find that further accelerative effects could enhance interdiffusion.
780: A specific mechanism might come into play which speeds up Pt particles
781: in the Ti bulk or at the interface.
782: The divergence of $\langle R^2 \rangle$ clearly indicates that accelerative effects are present
783: in the lattice.
784: At low ion energies, such as 0.5 keV, the damage energy
785: caused by the slowed down ions cannot accelerate particles beyond the
786: ion penetration depth.
787: Moreover, the observed asymmetry also could not be understood within a simple
788: picture of collisional cascades.
789: The mean diffusional depth could reach few tens of $\hbox{\AA}$, which
790: is beyond the mean free path of recoils at such a low ion energies.
791: We support the explanation raised in ref. \cite{Abrasonis} that nonlinear
792: forces could govern the accelereration of the particles in the bulk.
793: However, it is unclear which possible scenario (breather-like solitons, anharmonic longitudinal vibrations, etc.) explain the superdiffusive mechanism \cite{Abrasonis}.
794: %The metastable supersaturation of Pt in Ti could also lead to the amplification of interdiffusion.
795:
796:
797:
798:
799:
800: {\scriptsize
801: This work is supported by the OTKA grant F037710
802: from the Hungarian Academy of Sciences.
803: We wish to thank to K. Nordlund
804: for helpful discussions and constant help.
805: The work has been performed partly under the project
806: HPC-EUROPA (RII3-CT-2003-506079) with the support of
807: the European Community using the supercomputing
808: facility at CINECA in Bologna.
809: The help of the NKFP project of
810: 3A/071/2004 is also acknowledged.
811: }
812:
813: \vspace{-0.7cm}
814:
815: \begin{thebibliography}{99}
816:
817: \bib{Michely}
818: T. Michely, J. Krug, {\em Island, Mounds and Atoms}, Springer (2004).
819:
820:
821: %\bib{Schukin}
822: %V.A. Schukin, N. N. Ledentsov, D. Bimberg, {\em Epitaxy of Nanostructures},
823: %Springer (2004).
824:
825: %\bib{IBAD}
826: %R. A. Gonz\'ales, F. Yubero, J. M. Sanz,
827: %{\em Low Energy Ion Assisted Film Growth}, Imperial College Press, London, (2003).
828:
829:
830: %\bib{Ramana}
831: %C. V. Ramana, {\em et al.}, Phys. Rev. Lett.,
832: %{\bf 90}, 066101, (2003).
833:
834: %\bib{Egelhoff}
835: %W. F. Egelhoff, {\em et al.},
836: %Prog. Surf. Sci., {\bf 67}, 355 (2001).
837:
838: %\bib{Gomez05}
839: %L. G\'omez, C. Slutzky, J. Ferr\'on, Phys. Rev. {\bf B71}, 233402 (2005).
840:
841:
842: %\bib{Odunuga}
843: %S. Odunuga, {\em et al.}, Phys. Rev. Lett. {\bf 95}, 045901 (2005).
844:
845: %\bib{Cai}
846: %M. Cai, {\em et al.},
847: %J. Appl. Phys., {\bf 95}, 1996 (2004).
848:
849: \bib{Abrasonis}
850: G. Abrasonis, W. M\"oller, X. X. Ma, Phys. Rev. Lett. {\bf 96}, 065901-1, (2006).
851:
852: \bib{Buchanan}
853: J. D. R. Buchanan, {\em et al.},
854: Phys. Rev. {\bf B66}, 104427 (2002).
855:
856: \bib{Parascandola}
857: S. Parascandola, D. L. Williamson, Appl. Phys. Lett., {\bf 76}, 2194 (2000).
858:
859: \bib{Delugas}
860: P. Delugas and V. Fiorentini,
861: Phys. Rev. {\bf B69}, 085203 (2004).
862:
863: \bib{Avasthi}
864: D. K. Avasthi, W. Assmann, A. Tripathi, S. K. Srivastava, S. Ghosh, F. Grüner, and M. Toulemonde
865: Phys. Rev. {\bf B68}, 153106 (2003).
866:
867: \bib{Melis}
868: C. Melis, G. M. Lopez, and V. Fiorentini
869: Appl. Phys. Lett. {\bf 85}, 4902 (2004).
870:
871: \bib{Solmi}
872: S. Solmi and M. Bersani,
873: J. Appl. Phys. {\bf 87}, 3696 (2000).
874:
875: \bib{Sparks}
876: D. R. Sparks, R. G. Chapman, and N. S. Alvi,
877: Appl. Phys. Lett. {\bf 49}, 525 (1986).
878:
879: \bib{Venezia}
880: V.C. Venezia, R. Duffy, L. Pelaz, M.J.P. Hopstaken, G.C.J. Maas, T. Dao, Y. Tamminga and P. Graat,
881: Materials Science and Engineering {\bf B124-125}, 245 (2005).
882:
883: \bib{Slijkerman}
884: W. F. J. Slijkerman, P. M. Zagwijn, J. F. van der Veen, G. F. A. van de Walle, and D. J. Gravesteijn,
885: J. Appl. Phys. {\bf 70}, 2111 (1991).
886:
887: \bib{Schwarz}
888: R. B. Schwarz and W. L. Johnson,
889: Phys. Rev. Lett. {\bf 51}, 415-418 (1983).
890:
891: \bib{Vogl}
892: G. Vogl, W. Miekeley, A. Heidemann, and W. Petry,
893: Phys. Rev. Lett. {\bf 53}, 934-937 (1984).
894:
895: \bib{Sule_condmat}
896: P. S\"ule, condmat/06
897:
898:
899: \bib{Sule_NIMB04}
900: P. S\"ule, M. Menyh\'ard, K. Nordlund, Nucl Instr. and Meth. in Phys. Res.,
901: {\bf B226}, 517 (2004), {\bf B211}, 524 (2003).
902:
903: \bib{Sule_PRB05}
904: P. S\"ule, M. Menyh\'ard, Phys. Rev., {\bf B71}, 113413 (2005).
905:
906: %\bib{Gomez}
907: %L. G\'omez, {\em et al.}, Phys. Rev. Lett., {\bf 84}, 4397 (2000).
908:
909: %\bib{Menyhard}
910: %A. Barna, {\em et al.}, Appl. Surf. Sci. {\bf 242} 375 (2005).
911:
912: \bib{Luo}
913: G. M. Luo, {\em et al.}, Phys. Rev. {\bf B64}, 104427-5 (2001).
914:
915: \bib{Sule_SUCI}
916: P. S\"ule, Surf. Sci., {\bf 585}, 170 (2005).
917:
918: \bib{Sule_NIMB04_2}
919: P. S\"ule, M. Menyh\'ard, K. Nordlund, Nucl Instr. and Meth. in Phys. Res.,
920: {\bf B222}, 525 (2004).
921:
922: \bib{Sule_condmat06b}
923: S\"ule, M. Menyh\'ard, L. K\'otis, J. L\'ab\'ar, W. F. Egelhoff Jr.,
924: www.xxx.lonl.gov/archive/condmat/0610227 (fig1a-1b are excluded due to size limit),
925: can also be downloaded
926: from www.mfa.kfki.hu/$\sim$sule/papers/\\ condmat0610227.pdf, submitted
927: to J. Appl. Phys.
928:
929: %\bib{Karolewski}
930: %M. A. Karolewski, Nucl. Instrum. Meth. in Phys. Res. {\bf 243}, 6 (2006).
931:
932: %\bib{Thijsse}
933: %B. J. Thijsse, T. P. C. Klaver and E. F. C. Haddeman,
934: %Appl. Surf. Sci., {\bf 231-232}, 29 (2004).
935:
936: %\bib{Hofmann}
937: %S. Hofmann,
938: %Prog. Surf. Sci., {\bf 36}, 35 (1991).
939:
940:
941:
942: %\bib{Gnaser}
943: %H. Gnaser, {\em Low-energy Ion Irradiation of Solid Surfaces},
944: %in Solid-State Physics, Springer (1999)
945:
946:
947: %\bib{Zhong}
948: %Y. Zhong, Y. Ashkenazy, K. Albe, R. S. Averback, J. Appl. Phys., {\bf 94}, 4432.
949: %(2003).
950:
951: %\bib{Hanson}
952: %D. E. Hanson, B. C. Stephens, C. Saravanan, J. D. Kress, J. Vac. Sci. Technol.
953: %{\bf A19}, 820. (2001).
954:
955:
956:
957:
958: %\bib{Martin}
959: %G. Martin, Phys. Rev. {\bf B30}, 1424 (1984),
960: %G. Martin, P. Bellon, Solid State Physics, {\bf 50}, 189 (1997),
961: %R. A. Enrique, P. Bellon, Phys. Rev. {\bf B70}, 224106 (2004).
962:
963: %\bib{Lysenko}
964: %O. V. Lysenko, {\em et al.}, Phys. Rev. Lett., {\bf 89}, 126102-1 (2002).
965:
966: %\bib{Delaire}
967: %O. Delaire, T. Swan-Wood, B. Fultz,
968: %Phys. Rev. Lett., {\bf 93}, 185704-1 (2004).
969:
970:
971: \bib{Stepanyuk1}
972: %V. S. Stepanyuk, {\em et al.},
973: V. S. Stepanyuk, D. V. Tsivline, D. I. Bazhanov, W. Hergert, and A. A. Katsnelson
974: Phys. Rev. {\bf B63}, 235406 (2001).
975:
976: \bib{Stepanyuk2}
977: H. L. Meyerheim, V. Stepanyuk, A. L. Klavsyuk, E. Soyka, J. Kirschner,
978: Phys. Rev. {\bf B72}, 113403 (2005).
979:
980: %\bib{Tersoff}
981: %J. Tersoff, Phys. Rev. Lett. {\bf 74}, 434-437 (1995).
982:
983: %\bib{Nastasi}
984: %M. Nastasi, J. W. Mayer, J. K. Hirvonen, {\em Ion-Solid Interactions: Fundamentals
985: %and Applications}, Cambridge, (1996).
986:
987:
988:
989: %\bib{TRIM}
990: %TRIDYN, FZR-317, W. M\"oller and M. Posselt, Forschungzentrum Rossendorf, 01314 Dresden, Germany.
991:
992: %\bib{MM_TRIM}
993: %M. Menyh\'ard, Surf. Interface Anal. {\bf 26} 1001 (1998).
994:
995: \bib{Nordlund_ref}
996: %K. Nordlund, Comput. Mater. Sci, {\bf 3}, 448. (1995).
997: K. Nordlund, J. Tarus, J. Keinonen, S.E. Donnelly, R.C. Birtcher, Nucl. Instr. and Meth. in Phys. Res. {\bf B206} 189 (2003),
998: K. Nordlund, {\em et al.}, Phys. Rev., {\bf B57} 7556. (1998),
999: C. G. Zimmermann, M. Yeadon, K. Nordlund, J. M. Gibson, and
1000: R. S. Averback, Phys. Rev. Lett. {\bf 83}, 1163 (1999).
1001:
1002: \bib{Allen}
1003: M. P. Allen, D. J. Tildesley, {\em Comupter Simulation of Liquids}, (Oxford Science Publications, Oxford 1989)
1004:
1005: \bib{EAM}
1006: Y. Mishin, p. 459., in Handbook of Materials Modeling, Springer (2005).
1007:
1008: \bib{Goyhenex}
1009: C. Goyhenex, {\em et al.}, Phys. Rev. {\bf B60}, 2781 (1999).
1010:
1011: \bib{Levanov}
1012: N. A. Levanov, {\em et al.}, Phys. Rev. {\bf B61}, 2230 (2000).
1013:
1014: \bib{ZBL}
1015: W. Eckstein, {\em Computer Simulation of Ion-Solid Interactions}, (Springer, Berlin 1991).
1016:
1017: \bib{CR}
1018: F. Cleri, G. Mazzone and V. Rosato, Phys. Rev. {\bf B47}, 14541 (1993).
1019:
1020: \bib{web}
1021: http://www.mfa.kfki.hu/$\sim$sule/animations/ptti.htm.
1022:
1023: %\bib{CuCo}
1024: %P. S\"ule, {\em et al.}, submitted, cond-mat/0605130
1025:
1026:
1027: \bib{Vergara}
1028: L.L. Vergara, I.Vaquila, J. Ferron, Appl. Surf. Sci. 151. 129 (1999).
1029:
1030: %\bib{Barna}
1031: %A. Barna, {\em et al.}, J. Vac. Sci. Tech. {\bf A21} 553 (2003).
1032:
1033: %\bib{MM}
1034: %M. Menyh\'ard, {\em et al.},
1035: %Appl. Surf. Sci. {\bf 180} 315 (2001).
1036:
1037: %\bib{Berg}
1038: %S. Berg, I. Katardjiev, Surf. Coat. Tech., {\bf 84}, 353 (1996).
1039:
1040: \bib{anomalous}
1041: R. Metzler, J. Klafter, Phys. Rep., {\bf 339}, 1. (2000),
1042: R. Guantes, J. L. Vega, S. Miret-Art\'es, Phys. Rev., {\bf B64}, 245415 (2001).
1043:
1044:
1045: \bib{Swanson}
1046: B. I. Swanson,
1047: J. A. Brozik, S. P. Love, G. F. Strouse, A. P. Shreve, A. R. Bishop, W.-Z. Wang, and M. I. Salkola
1048: Phys. Rev. Lett. {\bf 82}, 3288-3291 (1999).
1049:
1050:
1051:
1052: \end{thebibliography}
1053:
1054:
1055: \end{document}
1056: