1: \documentstyle[aps,floats,graphicx]{revtex}
2:
3: \begin{document}
4:
5: \newcommand{\bec}{\begin{center}}
6: \newcommand{\ec}{\end{center}}
7: \newcommand{\be}{\begin{equation}}
8: \newcommand{\ee}{\end{equation}}
9: \newcommand{\beqn}{\begin{eqnarray}}
10: \newcommand{\eeqn}{\end{eqnarray}}
11: \newcommand{\bet}{\begin{table}}
12: \newcommand{\ent}{\end{table}}
13: \newcommand{\bib}{\bibitem}
14:
15:
16: %\baselineskip 4.2mm
17:
18: \wideabs{
19:
20: \title{
21: %What is the real driving force of ion beam mixing?
22: Strong mass effect on ion beam mixing in metal bilayers
23: %\begin{center}
24: %\end{center} Melting behavior and phase stability of ion bombarded metal bilayers: a ballistic picture
25: }
26:
27:
28: \author{P. S\"ule, M. Menyh\'ard}
29: \address{Research Institute for Technical Physics and Material Science,\\
30: Konkoly Thege u. 29-33, Budapest, Hungary, sule@mfa.kfki.hu,www.mfa.kfki.hu/$\sim$sule\\
31: %Accelerator Lab., Helsinki, Finland
32: }
33:
34: \date{\today}
35:
36: \maketitle
37:
38: \begin{abstract}
39: Molecular dynamics simulations have been used to study the mechanism of
40: ion beam mixing in metal bilayers.
41: We are able to explain the ion induced low-temperature phase stability and melting behavior of bilayers using only
42: a simple ballistic picture up to 10 keV ion energies.
43: The atomic mass ratio of the overlayer and the substrate constituents
44: seems to be a key quantity in understanding atomic mixing.
45: The critical bilayer mass ratio of $\delta < 0.33$ is required for the
46: occurrence of a thermal spike (local melting) with a lifetime of $\tau > 0.3$ ps at low-energy
47: ion irradiation ($1$ keV) due to a ballistic mechanism.
48: The existing experimental data follow the same trend as the simulated values.
49: \\
50:
51:
52: {\em PACS numbers:} 79.20.Rf, 61.80.Az 61.80.Jh 61.82.Bg\\
53: %\scriptsize{
54: %61.80.Az Theory and models of radiation effects,
55: %61.80.Jh Ion radiation effects,
56: %61.82.-d Radiation effects on specific materials,
57: %61.82.Bg Metals and alloys,
58: %66.30.-h Diffusion in solids (for surface and interface diffusion, see 68.35.Fx)\\
59: %79.20.Rf Atomic, molecular, and ion beam impact and interactions with surfaces\\
60: %81.40.Wx Radiation treatment (particle and electromagnetic) (see also 61.80.-x Physical radiation effects, radiation damage)\\
61: %}
62:
63: %{\em Keywords: Computer simulations, Ion-solid interaction, ion-beam mixing, molecular dynamics, mass effect, interfacial mixing, atomic migration,
64: %amorphisation}
65: \end{abstract}
66: }
67:
68:
69: %\section{Introduction}
70:
71: The mechanism of atomic transport in solids is a fundamental subject of material science
72: \cite{Allnatt}.
73: The atomic relocations, such as atomic mixing at interfaces under the effect of ion irradiation
74: has also been an important topic in the last decades \cite{Bellon,Okamoto}.
75: In these phenomena the atomic systems are driven far from equilibrium and the
76: roughening of the interface acts in competition with the restoring forces of the relaxation
77: process.
78: In certain interfacial systems ion beam mixing (IM) or the elevated temperature seems to randomize the initially sharp interface
79: leading to amorphization \cite{Samwer,Gyulai}, to crystal-glass \cite{Okamoto} or liquid to glass transition \cite{Faupel}. In many bilayers no considerable IM occurs \cite{mix_exp1,mix_exp}.
80: As an explanation for the ion induced phase stability (instability) of interfaces, the thermal spike
81: (TS) model is widely accepted in the last two decades \cite{mix_exp,AverbackRubia}
82: which predicts the dependence of the mixing efficiency on thermodynamic quantities such as
83: heat of mixing.
84: Recently though we observed the presence of TS but
85: failed to find any effect of
86: heat of mixing on IM \cite{Sule1,Sule3}.
87: Thus accepting the importance of the TS we propose to understand its affect on IM as
88: a purely ballistic phenomenon \cite{Sule3}.
89:
90:
91:
92:
93: % The ballistic mechanism has long been proposed for understanding IM \cite{Bellon,Sigmund},
94: %however, until nowdays never reached the widely accepted level due its apparent limitations,
95: %such as the too low predicted range of mixing \cite{AverbackRubia} when compared with the
96: %experiment \cite{mix_exp}.
97: %The reported limitations of the ballistic collision approximation model is based on the expectation that
98: %the ballistic model affects IM during a very short time period ($0.3$ ps, cascade mixing).
99: % However, it became apparent now that at least in certain metal bilayers cascade mixing
100: %persists up to several tens of ps \cite{Sule3}.
101: %%Moreover there are a couple of other studies which support the strong kinematic and ballistic
102: %effects under irradiation of solids \cite{Bellon}.
103:
104:
105: In the present Letter we would like to show that
106: the atomic mass ratio has a dramatic effect on
107: IM in various metal bilayers. At 1 keV Ar$^+$ ion bombardment no real TS occurs above the atomic mass ratio of
108: $\delta=m_{overl}/m_{bulk}>0.33$ while considerable IM occurs if $\delta < 0.33$, where $m_{overl}$ and $m_{bulk}$ are the atomic masses in the overlayer and in the substrate (bulk).
109: We will show that the measured mixing efficiencies follow the same trend as the simulated
110: IM.
111: We propose to understand IM as a ballistic process and we find that the chemical
112: interdiffusion model \cite{Cheng} might be inappropriate.
113:
114:
115:
116:
117:
118: %\section{The setup of the simulation}
119:
120: Classical constant volume molecular dynamics simulations were used to simulate the ion-solid interaction
121: using the PARCAS code \cite{Nordlund_ref}.
122: The computer animations can be seen in our web page \cite{web}.
123: Here we only shortly summarize the most important aspects.
124: Its detailed description is given in \cite{Nordlund_ref} and details specific to the current systems in recent
125: communications \cite{Sule1,Sule2}.
126: We irradiate a series of bilayers (Al/Pt, Ti/Pt, Al/Ag, Cu/Pt, Cu/Au, Al/Cu, Ni/Ag, Ag/Au, Ti/Co, Ag/Ni,
127: Cu/Ni) with 1 keV Ar$^+$ ions. In certain bilayers (Al/Pt, Ti/Pt, Al/Cu, Ni/Ag, Ti/Co) we
128: increase the ion energy up to 10 keV in order to show that the mass effect
129: persists up to higher ion energies.
130: The initial velocity direction of the
131: impacting atom was $7$ degrees with respect to the surface of the crystal (grazing angle of incidence).
132: To obtain a representative {\em statistics},
133: the impact position of the incoming ion is varied
134: randomly within a $5 \times 5$ \hbox{\AA}$^2$ area.
135:
136: We used a tight binding many body potential given by Cleri and
137: Rosato (CR) \cite{CR}, to describe interatomic interactions.
138: We have chosen those bilayers for which atomic potentials are available.
139: This type of a potential gives a very good description of lattice vacancies, including migration
140: properties and a reasonable description of solid surfaces and melting \cite{CR}.
141: %Since the present work is mostly associated with the elastic properties,
142: %melting behaviors, surface, interface and migration energies, we believe the model used should be
143: %suitable for this study.
144: The bilayer AB cross potential is composed of the CR potentials of the A and B elements.
145: Taking the simple average of the elemental potentials the heat of mixing is set to $\Delta H_m \approx 0$.
146: This can be done since
147: we have shown recently, that
148: we found no considerable dependence of IM on $\Delta H_m$ in Ti/Pt \cite{Sule1} up to
149: 7 keV ion energy \cite{Sule3}.
150: The choice of $\Delta H_m \approx 0$ is also useful from that point of view that
151: we would like to show that even in the case of $\Delta H_m \approx 0$ (the lack of AB chemical bonding) considerable
152: IM occurs in certain bilayers.
153: In other bilayers the effect of $\Delta H_m$
154: on IM is also very weak if any \cite{Sule3}.
155: The reliability of the AB crosspotential as an average of the elemental potentials is tested
156: in the case of CuAu, where an optimized CR potential is available for the alloy phase
157: \cite{CR2}. We find no considerable changes in the physical properties or in the equilibrium
158: structure. Although this does not prove the reliability of other crosspotentials,
159: it suggests that the average potentials are very close to the optimized one.
160:
161:
162: The construction of the interface systems is given elsewhere \cite{Sule3}.
163: We only shortly summarize that
164: the interfaces have (111) orientation and the
165: close packed directions are parallel.
166: %------------------------------------------------------
167: \begin{figure}[hbtp]
168: \begin{center}
169: \includegraphics*[height=4.5cm,width=6cm,angle=0.0]{mix_massr_mod.eps}
170: \caption[]{
171: The number of intermixed atoms ($N_{mix}$, interfacial mixing) at various atomic mass ratio ($\delta$) in various
172: metal bilayers obtained at 1 keV Ar$^+$ irradiation.
173: Note the threshold $\delta \approx 0.33$ value below which
174: the bilayers exhibit an ion-beam mixing "catastrophe" and above which
175: only cascade mixing occurs.
176: The data points correspond from left to right to the bilayers Al/Pt (0.14), Al/Ag (0.24), Ti/Pt (
177: 0.25), Cu/Au (0.32), Cu/Pt (0.33), Al/Cu (0.42), Ni/Ag (0.54), Au/Ag (1.83), Ti/Co (0.81), Cu/Ni (1.08) and
178: Ag/Ni (1.82), respectively
179: (the atomic mass ratios are given in the paranthesis).
180: In those case where $\delta > 1$ we use $1/\delta$, because we find that $N_{mix}$ in AB and BA
181: systems
182: are nearly equal.
183: %The dashed line is for guiding the eyes.
184: The error bars denote standard deviations.
185: {\em Inset}: The number of mixed atoms is also shown for higher energies for
186: Al/Pt (6 keV), Ti/Pt (8 keV), Al/Cu (10 keV), Ni/Ag (9 keV) and for Ti/Co (10 keV) as a function
187: of the mass ratio.
188: }
189: \label{mass}
190: \end{center}
191: \end{figure}
192: %------------------------------------------------------
193: The thickness of the upper layer is 4-8 monolayer (ML), while the bulk is constructed
194: from 36 MLs in those samples which subjected to 1 keV irradiation.
195: These samples includes roughly 45000 atoms.
196: At higher irradiation energy we put a thicker overlayer with 8-16 MLs and a substrate
197: with 90 MLs (550000 atoms).
198: The interfacial systems
199: are created as follows:
200: the overlayer is put by hand on the (111) substrate (bulk) and various structures are probed
201: and are put together randomly. Finally that one is selected which has the smallest
202: misfit strain prior to the relaxation run.
203: The remaining misfit is properly minimized during the relaxation
204: process so that the overlayer and the substrate layers keep their original crystal structure and we get an
205: atomically sharp interface.
206: %The layers at the interface initially are separated by $\sim 2.8$ {\AA} and allowed freely to relax during the simulations.
207:
208:
209: We carry out {\em liquid and hot atom analysis} in order to study the role of highly mobile particles
210: in IM \cite{Sule3}.
211: To do so we need to calculate the
212: time dependent temperature of the $i$th high energy particle ($T_{i,local}(t)$) which can be given as follows:
213: $\frac{1}{2} m_i v_i^2(t) = \frac{3}{2} k T_{i,local}(t)$
214: where $m_i$ and $v_i$ are the atomic mass and the velocity of the $i$th high energy particle and $k$ is the Boltzmann constant.
215: We allow
216: some temperature fluctuation for the liquid atoms above $T_m$ (the melting point), hence for a hot atom and for recoils we use the arbitrary definition
217: $T_m+1000 < T_{local}$ \cite{Sule3}.
218: Hot atoms are those energetic particles which have a medium long mean path
219: ($\sim 10$ \hbox{\AA})
220: with an average kinetic energy of 0.1-1 eV.
221: %These kinetic energy values are in the range of or higher than
222: %the heat of mixing ($\sim 0.1-1$ eV/atom) \cite{Sule3}.
223: %------------------------------------------------------
224: \begin{figure}[hbtp]
225: \begin{center}
226: \includegraphics*[height=4.5cm,width=6cm,angle=0.0]{mixeff_massrat_expcalc.eps}
227: \caption[]{
228: The low-temperature experimental mixing efficiency (\hbox{\AA}$^5$/eV) as a function of the
229: atomic mass ratio ($\delta$). The values are taken from refs. \cite{mix_exp1,mix_exp}.
230: In those cases where $\delta > 1$ we use instead $1/\delta$ because we find that there is no
231: serious difference in mixing between AB and BA bilayers.
232: {\em Inset}: The calculated (simulated) $\xi_{IM}$ (\hbox{\AA}$^5$/eV) vs. $\delta$.
233: }
234: \label{exp_mass}
235: \end{center}
236: \end{figure}
237: %------------------------------------------------------
238: Liquid atoms have shorter, recoils have much longer mean free path.
239: A liquid atom can be considered as a nearly standing atom during the lifetime of the TS and only vibrationally
240: excited around its equilibrium position while a hot atom jumps to another position with
241: a velocity of $v_i \approx 1$ \hbox{\AA}/ps \cite{Sule3}.
242: %The mobile species have a limited interaction with the lattice due to the high mobility of them.
243: %We found a considerable amount of hot atoms in various bilayer at low energy irradiation persisting up to several
244: %ps in a previous study \cite{Sule3}.
245: %Interesting result of this article is that the lifetime of the cascade period
246: %is much longer in metal bilayers with the atomic mass ratio of $\delta < 0.33$ then usual $\tau_{recoil} < 3-500$ Fs \cite{AverbackRubia}.
247: The hot atoms are present till the end of the TS hence cascade and TS
248: mixing coexists in certain interfacial systems.
249: The more precise definition of a hot atom is given elsewhere \cite{Sule3}.
250:
251:
252: We calculate the number of mixed atoms ($N_{mix}$) and the simulated mixing efficiency $\xi_{IM}^{sim}=\frac{\langle R^2 \rangle}{6 n_0 E_{D_N}}$, where $\langle R^2 \rangle$, $n_0$ and $E_{D_N}$,
253: are the calculated mean square atomic displacement through the interface per atom, the atomic density in the
254: upper layer and the deposited nuclear energy.
255: %with the available high energy values \cite{mix_exp1} which are obtained by the
256: %measurement of broadening at the interface.
257: We exclude from $\langle R^2 \rangle$ atomic displacements which do not lead to
258: broadening at the interface (self-atomic mixing) because the experimental $\xi_{IM}$ is
259: calculated from broadening at the interface \cite{mix_exp,AverbackRubia}.
260: Further calculational details are given in ref. \cite{NordlundNIMB00}.
261:
262:
263: We investigate the influence of the atomic mass ratio on ion beam mixing.
264: We ion bombarded various bilayer systems with different atomic mass ratios and
265: find that below a threshold ratio ($\delta < 0.33$) the magnitude of
266: intermixing ($N_{mix}$) is enhanced abruptly.
267: The results are summarized in FIGs (1)-(2).
268: In FIG (1) we plot the simulated $N_{mix}$ vs. $\delta$ and
269: the experimental $\xi_{IM}$ vs. $\delta$ in FIG (2) collected from refs. \cite{mix_exp1,mix_exp}.
270: In the inset FIG (2) we also give the simulated $\xi_{IM}$ vs. $\delta$
271: for bilayers for which high energy (up to 10 keV) simulations are available and also for
272: the other bilayers for which $\xi_{IM}^{sim}$ is calculated at 1 keV ion energy.
273: %Most experimental irradiations have been performed with heavy ions ($Xe^+$) and with much higher
274: %ion energies ($100-600$ keV) than the ones used here.
275: %Therefore we prefer to use $N_{mix}$ vs. $\delta$ plot which nicely reflect the experimental
276: %trends given in FIG (2).
277: In FIG (2) wee see that the increase in $\xi_{IM}$ occurs between $\delta \approx 1/3$
278: and $1/2$ which is rather similar to the inset FIG (1).
279: %------------------------------------------------------
280: \begin{figure}[hbtp]
281: \begin{center}
282: \includegraphics*[height=4.5cm,width=6cm,angle=0.0]{tipt_1kev_recoil_xz_10A.eps}
283: \caption[]{
284: The $xz$ atomic positions of the recoils and the hot atoms at 1 keV ion energy collected up to $4
285: $ ps in Ti/Pt.
286: The z-coordinate is the depth position.
287: The dashed line denotes the interface.
288: %a typical intermixed Pt recoils. Their lifetime is around 3 ps.
289: }
290: \label{cascade1}
291: \end{center}
292: \end{figure}
293: %------------------------------------------------------
294: In FIG ~(\ref{exp_mass}) the experimental mixing efficiency values are obtained as follows \cite{mix_exp1}: ($\xi_{IM} \approx \frac{D t}{\Phi F_D}$), where $D t$ is the diffusion length, $\Phi$ is the fluence and $F_D$ is the deposited energy at the interface).
295: Although there is some scatter in the data
296: at around $\delta \approx 1/3$, $\xi_{IM}$ increases heavily in accordance with the
297: $N_{mix}$ values shown in FIG ~(\ref{mass}).
298: Although the values in FIG ~(2) are obtained at high energies, however,
299: they can be compared with our $\xi_{IM}^{sim}$ values since IM occurs
300: primarily in the subcascade region, where the energy is in the range we applied \cite{NordlundNIMB00}.
301:
302:
303:
304: In particular in Al/Pt we find a strong amorphization and broadening at the interface
305: which is in accordance with measurements \cite{Samwer,Gyulai,Sule3,Hung}.
306: In Cu/Pt and in Cu/Au the phase stability of the interface is also very weak
307: in accordance with ion irradiation experiments \cite{Hung}.
308: In Cu/Au 1 MeV ion bombardment results in strong broadening \cite{mix_exp,Hung}.
309: %Even in the case of annealing of the Cu/Au bilayer at 450 K a strong interdiffusion is
310: %observed \cite{Hartung}.
311: In bilayers Al/Ag and in Au/Ag we find a relatively weak interfacial mixing ($N_{mix} < 30$),
312: however, $\xi_{IM}^{sim}$ is large ($\xi_{IM}^{sim} \approx 300 \pm 120$ in Al/Ag,
313: $\xi_{IM}^{sim} \approx 115 \pm 45$ in Au/Ag).
314: A large measured value of $\xi_{IM}^{exp} approx 265$ is found in Au/Ag
315: \cite{mix_exp1} and this data point is plotted at $1/\delta=0.55$ in FIG (2).
316: Interestingly the atomic mobility is relatively large in these systems while the gross number of
317: mixed atoms ($N_{mix}$) is small.
318: We attribute this "anomalous" behavior of Au/Ag to the tendency of crater formation in Au \cite{AverbackRubia}.
319: It has been shown, recently,
320: that crater formation enhances mass transport hence atomic mobility between the interface and the free surface \cite{Sule2}.
321: %\begin{table}
322: %\caption[]
323: %{
324: %}
325: %{\scriptsize
326: %\begin{tabular}{cccccc}
327: % A/B & \delta & $\xi_{IM}^{calc}$ & $\xi_{IM}^{calc}$ & $\Delta \sigma & \\
328: %\hline
329: % Al/Pt & 0.14 & 324 & 25 & \\
330: % Ti/Pt & 0.24 & 2.0 & 9 & \\
331: % Al/Cu & 0.42 & 1.6 & 6 & \\
332: % Ni/Ag & 0.54 & 0.8 & 3 & \\
333: %Ti/Co & 0.82 & 0.5 & 3 & \\
334: %---------------------------------------------------------------
335: %\end{tabular}}
336: %{\scriptsize
337: %}
338: %\label{}
339: %\end{table}
340: %We find the following values for broadening $\Delta \sigma$ (\hbox{\AA}) at the interface
341: %after 1 keV irradiation (values for higher energies are in paranthesis)
342: %Al/Pt: $\sim 16$ (6 keV: $\sim 25$), Cu/Au: $\sim 12$, Ti/Pt: $\sim 9$, Al/Cu $\sim 3 (10 keV:6)$, Ni/Ag: $\sim 3 (9 keV: 6)$.
343:
344:
345:
346: In FIG ~(\ref{exp_mass}) the experimental mixing efficiency values \cite{mix_exp1,mix_exp} ($\xi_{IM} \approx \frac{D t}{\Phi F_D}$), where $D t$ is the diffusion length, $\Phi$ is the fluence and $F_D$ is the deposited energy at the interface) are plotted against
347: the atomic mass ratio.
348: Although there is some scatter in the data
349: at around $\delta \approx 1/3$, $\xi_{IM}$ increases heavily in accordance with the
350: $N_{mix}$ values shown in FIG ~(\ref{mass}).
351: Although the values in FIG ~(2) are obtained at high energies, however,
352: they can be compared with our $\xi_{IM}^{sim}$ values since IM occurs
353: primarily in the subcascade region, where the energy is in the range we applied \cite{NordlundNIMB00}.
354:
355:
356:
357:
358: For bilayers with $\delta > 0.33$ starting with Al/Cu ($\delta = 0.42$) we got a weak interfacial mixing in accordance
359: with the MD \cite{Colla,NordlundPRB99} and experimental studies \cite{mix_exp1}.
360: The strong mass effect can be understood as the ballistic mechanism
361: plays an essential role in IM and
362: may stem
363: from an increased backscattering of light overlayer atoms from the
364: heavy substrate \cite{Sule3}.
365: %------------------------------------------------------
366: \begin{figure}[hbtp]
367: \begin{center}
368: \includegraphics*[height=4cm,width=5.5cm,angle=0.0]{tico_1kev_recoil_xz_10A.eps}
369: \caption[]{
370: The $xz$ positions of the recoils and the hot atoms at 1 keV ion energy collected up to $0.3$ ps in Ti/Co.
371: The z-coordinate is the depth position.
372: The dashed line denotes the interface.
373: %a typical intermixed Pt recoils. Their lifetime is around 3 ps.
374: }
375: \label{cascade2}
376: \end{center}
377: \end{figure}
378: %------------------------------------------------------
379: Moreover the density of the collisional cascades depends on the bilayer mass ratio \cite{Sule3}.
380: It should also be stressed that in pure elements or even in metal
381: alloys we find a much weaker IM and shorter TS \cite{web,Sule2} due to
382: the strong (re)ordering forces.
383: In order to elucidate the relative roles of ballistics and TS in the
384: interfacial mixing
385: we examined the prototypical bilayer Ni/Ag ($\delta \approx 0.54$) and
386: found a very weak interfacial mixing at 1 keV bombardment and no occurrence of a real TS period.
387: The energetic atoms (recoils, hot atoms) dissappear at less then 0.3 ps,
388: which is typically the end of the cascade period.
389: This system is relatively well studied
390: theoretically \cite{Colla} at 10 keV ion energy and being a typical
391: example of a segregating system which has a high positive heat of mixing
392: in the liquid \cite{Colla}.
393: At 9 keV we find a real thermal spike which persists up to 3 ps.
394: The number of mixed atoms ($N_{mix}$) is, however, much smaller then in Al/Pt or in Ti/Pt (inset FIG ~\ref{mass})).
395: Therefore, although there is an increase in $N_{mix}$, the trend remains the same: the effect of
396: the mass ratio is also robust at higher energies.
397: We attribute, however, the weak IM in Ni/Ag not to the positive heat of mixing, but to the weak mass
398: effect in this system ($0.33 < \delta \approx 0.54$).
399: In Ti/Co we find no TS up to 10 keV as well as IM is very weak in this system ($\delta \approx 0.82$).
400: It should be emphasized that the similar situation is true for all the bilayer samples
401: which have $\delta > 0.33$ (FIG ~(\ref{mass})).
402: This observation clearly indicates a robust mass effect when $\delta < 0.33$, hence a kinematic and ballistic
403: picture seems to be sufficient for describing IM in bilayer systems.
404: In those systems, where TS does not occur, cascade mixing is the only IM effect.
405: %In our previous reports we have also presented that
406: %the TS has a considerable effect on mixing, though not in the chemically biased
407: %way as it is known in the literature \cite{Sule1,Sule3}.
408: %This is in contrast with an earlier study, where the effect of heat of mixing is
409: %explored in a model system \cite{Colla}.
410: %In ref. \cite{Sule3} we discuss the possible reason of this discrepancy.
411:
412:
413:
414:
415: The cascade events during the ballistic period is shown in FIGs ~(\ref{cascade1})-
416: (\ref{cascade2}).
417: In the case of Ti/Pt we get a dense collision cascade, the recoils (hot atoms) are concentrated
418: within a
419: smaller region due to couple of reflections while in Ti/Co the high energy particles
420: are scattered in a larger volume hence the deposited energy spreads over a larger irradiated
421: region. Therefore if the mass ratio $\delta > 0.33$, the cooling of the cascade
422: is ultrafast due to the low concentration of the recoils (hot atoms).
423: Indeed, we found that the average atomic concentration of the hot atoms in the irradiated zone ($V \approx 1000 \hbox{\AA}^3$) is
424: around $10^{22}$ atom/cm$^3$ for $\delta > 0.33$ and $4 \times 10^{22}$ atom/cm$^3$ for $\delta <
425: 0.33$.
426: These values should also be compared with the average atomic concentration
427: of $5-8 \times 10^{22}$ atom/cm$^3$ in metals.
428: If the mass ratio drops below $0.33$, the hot particles are still present in the TS.
429: One can see that for $\delta < 0.33$ the peak hot atom concentration is close to the
430: atomic concentration. In these systems we no longer have a simple liquid ensemble, it is rather
431: a superheated system \cite{Sule3}.
432: The mass effect is even more pronounced at higher energies. The corresponding plots of
433: the collisional cascades in Al/Pt at 6 keV ion energy can be seen in a web page \cite{web2}.
434: We have also shown that the density of the cascade is strongly affected by $\delta$ \cite{Sule3,web}.
435:
436:
437:
438: {\em Qualitatively we explain the observed strong mass effect as follows:}
439: In elastic collisions of the recoils (energetic light particles from the overlayer) with the heavier substrate atoms
440: the kinetic energy of the moving atoms is partly transferred to the heavier atoms, which
441: , however, might not be kicked out of their positions because of the large mass difference.
442: The colliding heavy partner of the recoil becomes vibrationally excited, which means that its rms amplitude
443: of thermal vibrations becomes equal to about 50 \% of the interatomic distance. That is basically
444: the Lindeman`s criterion for lattice instability during melting: a crystal melts when the rms thermal
445: displacement of atoms from their equilibrium positions become large enough to invade their
446: nearest-neighbor spaces \cite{Okamoto}.
447: For such thermal displacements, the thermal expansion would far exceed the critical value for
448: shear instability (the Born criterion, the lost of at least one of the shear moduli) leading to mechanical instability \cite{Okamoto}.
449: The neighborhood of this hot heavy atom is heated up and local melting
450: occurs (TS).
451: When the mass of a recoil and the colliding partner is comparable the target atom
452: might be displaced from its original position leaving a vacant site. In this case
453: the slowing down of the recoil in the bulk does not result in local melting because the
454: kinetic energy of the recoil spreads over a too large volume.
455:
456:
457: The important question remained to be answered what is the reason of the critical mass ratio
458: of $\delta \approx 0.33$?
459: We do believe that the threshold value is due to the emergence of a strong backscattering of the recoiling light atoms at the interface. Below
460: this value hence the energy deposition becomes extremely effective at the interface through energy
461: transfer to the standing heavy atoms.
462: The backscattering effect at the interface results in the confinement of the light recoils
463: in the overlayer which leads to superheating. This is the primary reason of the high concentration
464: of hot atoms in these bilayers (see FIGs ~(\ref{cascade1})).
465: The interfacial backscattering phenomenon can be attributed partly to the mass difference
466: and also to other effects such as the difference in the cohesive energies in the
467: substrate and in the overlayer \cite{Sule3}.
468:
469:
470:
471: In summary, we have shown that intermixing in metal bilayers strongly depends
472: on the relative masses of the constituents under the effect of ion irradiation.
473: There exists a threshold mass ratio value below which the interface system is unstable
474: against ion bombardment.
475: We propose to understand ion beam mixing as a ballistic process.
476: The observed strong mass effect in heterophases might be an important topic in preparation
477: of thin films and multilayers especially with great technological importance.
478:
479:
480:
481:
482:
483:
484:
485:
486:
487: %\section{acknowledgment}
488: {\small
489: This work is supported by the OTKA grant F037710
490: from the Hungarian Academy of Sciences. We greatly
491: acknowledge conversations with K. Nordlund.}
492:
493:
494:
495:
496: %\section{Conclusion}
497:
498: \vspace{-0.5cm}
499:
500: \begin{thebibliography}{99}
501:
502: \vspace{-1.5cm}
503:
504: \bib{Allnatt}
505: A. R. Allnatt, A. B. Lidiard, {\em Atomic Transport in Solids}, Cambridge
506: University
507: Press, Cambridge, England, (1993)
508:
509: \bib{Bellon}
510: G. Martin, P. Bellon, Solid State Phys., {\bf 50}, 189 (1997).
511:
512: \bib{Okamoto}
513: P. R. Okamoto, N. Q. Lam, L. E. Rehn, Solid State
514: Phys., {\bf 52}, 1 (1999).
515:
516:
517: %\bib{Gnaser}
518: %H. Gnaser, {\em Low-Energy Ion Irradiation of Solid Surfaces}, Solid-State Physics, {\bf 146} (1999), Springer.
519:
520:
521: \bib{Samwer}
522: K. Samwer, H. J. Fecht and W. L. Johnson, {\em Amorphization in Metallic Systems},
523: in Glassy Metals III, eds. H. Beck, H.-J. Güntherodt, Springer, (1994).
524:
525: \bib{Gyulai}
526: L. S. Hung, M. Nastasi, J. Gyulai, and J. W. Mayer, Appl. Phys. Lett., {\bf 42},
527: 672 (1983).
528:
529: \bib{Faupel}
530: F. Faupel, W. Frank, M. Macht, H. Mehrer, V. Naundorf, K. R\"atzke, H. R. Schober, S. K.
531: Sharma, H. Techler, Rev. Mod. Phys., {\bf 75}, 237 (2003).
532:
533: \bib{mix_exp1}
534: B. M. Paine, R. S. Averback, Nucl. Intsr. Meth. {\bf B7/8}, 666 (1985).
535:
536: \bib{mix_exp}
537: %A. Crespo-Sosa, M. Munoz, J.-C. Cheang-Wong, A. Oliver, J. M. Saniger, J. G. Banuelos, Mat. Sci. Eng. {\bf B100}, 297 (2003),
538: W. Bolse, Mat. Sci. Eng. {\bf A253}, 194 (1998),
539: L. C. Wei, R. S. Averback, J. Appl. Phys. {\bf 81}, 613 (1997),
540: %T. Weber, K. Lieb, J. Appl. Phys., {\bf 73}, 3499 (1993),
541: S.-J. Kim, M-A. Nicolet, R. S. Averback, D. Peak, Phys. Rev. {\bf B37}, 38 (1985),
542: %T. A. Workman, Y. T. Cheng, W. L. Johnson, and M.-A. Nicolet, Appl. Phys. Lett. {\bf 50}, 1486 (1987).
543:
544:
545: \bib{AverbackRubia}
546: R. S. Averback, T. Diaz de la Rubia, Solid State Pysics, {\bf 51}, 281 (1998), and references therein.
547:
548: \bib{Sule1}
549: P. S\"ule, M. Menyh\'ard, K. Nordlund, Nucl Instr. and Meth. {\bf B211}, 524 (2003),
550: http://xxx.lanl.gov/cond-mat/0302262.
551:
552: \bib{Sule3}
553: P. S\"ule, M. Menyh\'ard, K. Nordlund, before submittion,
554: http://xxx.lanl.gov/cond-mat/0405280.
555:
556: \bib{Cheng}
557: Y.-T. Cheng, Material Science Reports, {\bf 5}, 45 (1990).
558:
559: \bib{Nordlund_ref}
560: K. Nordlund, M. Ghaly, R. S. Averback, M. Caturla, T. Diaz de la Rubia, and J. Tarus, Phys. Rev. {\bf B57}, 7556 (1998).
561:
562: \bib{web}
563: P. S\"ule, http://www.mfa.kfki.hu/$\sim$sule (2004).
564:
565: \bib{Sule2}
566: P. S\"ule, M. Menyh\'ard, K. Nordlund, Nucl Instr. and Meth. B, in press,
567: http://xxx.lanl.gov/cond-mat/0310238.
568:
569: \bib{CR}
570: F. Cleri and V. Rosato, Phys. Rev. {\bf B48}, 22 (1993).
571:
572: \bib{CR2}
573: F. Cleri, G. Mazzone and V. Rosato, Phys. Rev. {\bf B47}, 14541 (1993).
574:
575: \bib{NordlundNIMB00}
576: K. Nordlund, J. Tarus, J. Keinonen, M. Ghaly, R. S. Averback, Nucl Instr. and Meth. {\bf B164-165}, 441 (2000).
577: K. Nordlund, M. Ghaly, R. S. Averback, J. Appl. Phys., {\bf 83}, 1238 (1998).
578:
579: \bib{Hung}
580: L. S. Hung, J. W. Mayer, Nucl Instr. and Meth., {\bf 7-8}, 676 (1985).
581:
582: %\bib{Hartung}
583: %F. Hartung, G. Schmitz, Phys. Rev. {\bf B64}, 245418-1 (2001).
584:
585: \bib{Colla}
586: T. J. Colla, H. M. Urbassek, Phys. Rev. {\bf B63}, 104206 (2001),
587: %\bib{Gades}
588: H. Gades, H. M. Urbassek, Phys. Rev. {\bf B51}, 14559 (1995).
589:
590: \bib{NordlundPRB99}
591: K. Nordlund, R. S. Averback, Phys. Rev. {\bf B59}, 20 (1999).
592:
593: \bib{web2}
594: http://www.mfa.kfki.hu/$\sim$sule/figs/cascade.htm.
595:
596:
597:
598: \end{thebibliography}
599:
600:
601:
602:
603: \end{document}
604: