HAL Id: jpa-00230794
https://hal.archives-ouvertes.fr/jpa-00230794
Submitted on 1 Jan 1990
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
INTERFACIAL ION MIXING IN METAL-SILICON BILAYERS
A. Ektessabi, A. Heshmati, R. Khosravi
To cite this version:
A. Ektessabi, A. Heshmati, R. Khosravi. INTERFACIAL ION MIXING IN METAL- SILICON BILAYERS. Journal de Physique Colloques, 1990, 51 (C4), pp.C4-281-C4-284.
�10.1051/jphyscol:1990434�. �jpa-00230794�
COLLOQUE DE PHYSIQUE
Colloque C4, suppl6ment au n014, Tome 51, 15 juillet 1990
INTERFACIAL ION MIXING I N METAL-SILICON BILAYERS
A.M. EKTESSABI, A.H. HESHMATI and R. KHOSRAVI
Ion Beam Application Laboratories, AEOI, PO. Box 11365-8486, Teheran, Iran
Abstract -
There has been a great deal of interest in the fabrication of new materials with unique properties using ion mixing technique. Ion mixing is also a strong tool for better understanding of ion- solid interactions such as, amorphisation process, and irradiation damage. In this regard, an attempt was made by this group to simulate the effects of energetic ion motion in metal-silicon bilayers by using a program TRIM-MIXING based on the program TRIM developed by Biersack et al. The main features of this version of the program are :-We have reached recoil profiles which are more accurate than what is presented in the literature by following the collisional cascades of short and long range recoil atoms throughout the sample.
-Mixing interface is more precisely characterized by taking into accoung both the recoils of the f i t layer into the second layer and vice versa. Simulations were performed on various Au-Si bilayers. The intermixing recoil profiles are presented and compared to other experimental and theoretical data available in the literature. Our data is in very good agreement with the experimental results obtained by other authors.
1 . Introduction
Ion beam induced mixing of metal on silicon bilayers has been used very effectively to creat new solid compounds during the past decade 11-53. Theoretical treatment of ion mixing process has also been the subject of a great deal of research as well [6-71. Qualitatively, ion mixing is a three step process. The prompt process, in,which the incident energetic ions trigger a series of atomic collisions and in the time order of 10-12seconds the discrete layers are mixed at their interface. This leaves the solid in an excited state with a high defect concentration at the mixing interface. The thermal spike process, where the energy deposited into the solid during the prompt process, causes a localized increase of temperature which spreads over the entire solid. At the time period in the order of 10-l1 seconds the solid relaxes to a state of lower free energy. The diffusion process, in which the radiation enhanced diffusion of defects created during the prompt process and moving of both metal and silicon atoms occur. This process requires longer time compared to the two other processes.
The last two processes are called the delayed process.
There is a lot of controversy on the mechanisms of mixing in different temperature regimes. It has been suggested that for metal-silicon systems, at low temperatures, the mixing becomes completely temperature independent [2,61 and mixing is controlled by the prompt process.
However, it was found that the broadening of the interface profiles at very low temperatures are larger than what was predicted by theory [8]. Also, it has been argued that the transport Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1990434
C4-282 COLLOQUE DE PHYSIQUE
mechanisms cannot just be the normal radiation-enhanced one but rather it is a rapid migration in the cascade region [g]. For example, in Pt-Si and Ni-Si [4-51 the dominant moving species at high temperature are metal atoms while at relatively lower temperatures Si atoms are the faster moving species.
There has been many attempts to develope simulation programs to predict the experimental result of ion mixing [10-111, but the agreement of experiment and theory is still poor. The aim of the present study is to develope a simulation program to obtain better theoretical data than what is presented in the literature by following the collisional cascades of the short and long range recoils in both sides of the mixing interface. Because of the complexity of the subject, in this work we have neglected the temperature dependence of mixing, Therefore, mixing effect is solely controlled by the prompt process.
2. Program Discription
The physical model used in our program called TRIM-MIXING is based on the monte car10 simulation program TRIM [121. To our knowledge until now the simulation programs developed to obtain recoil profiles use the following logic: As an ion enters the top layer, in a distance of one mean free path it collides with an atom. If the energy transfer is greater than the binding energy, the atom recoils at a certain angle different than the deflection angle of the incident ion. The recoiled atom may collide with another atom creating recoils of itself until the generation of successive recoils stops. At this point the complete path of the recoil generation is known. The program now goes back to the atom which created the last recoil and follows its path until it comes to rest. This process continues up to where the first recoil was generated. Then the program follows the ion path and repeats the same procedure.
TRIM-MIXING gives more accurate recoil profiles due to the following modifications: (a) in contrast to other programs we have taken different values of mean free path for ion and recoil atoms before and after each collision. In addition, for ion-atom collisions and recoil-atom collisions, we have calculated different values of deflection angles, reduce mass and collision cross sections. (b) Provisions to follow ion and recoil paths in bilayared system.
The method used is to simply identify each incoming ion, and recoils of each layer across the interface by a given code. In this way, the profiles of the stopped recoils in each layer versus depth as well as the thickness of the mixed layer can be obtained. (c) If only implanted recoils of the first layer into the second layer is of interest, by omitting the soft collisions that generates recoils which never reach the interface, we have considerably reduce the run time of the program.
3. Results and Discussion
We have chosen to simulate mixing of Au-Si bilayer using 4r, Kr and Xe ions similar to the experiments done by Poprocki et.al [13]. The simulations were performed assuming a 25-30nm thick Au layer on silicon and mixing the bila er using 120 kev Ar ions at the d o s e d 5 ~ ~ 0 1 4
ions/cm2
,
200 kev Kr ions at the dose of .5x10 ions/cm2 T4,
and 300 kel Xe ions at the dose of 1 0 ~ ~ i o n s / c m ~.
For comparison, Figure 1. shows the results of TRIM-MIXING with the experimental data obtained by Poprocki et.al 1131. The results from a computer simulation used by Poprockj et.al and calculations usins Gras-Marti analytical formulae [14-151 are also included in these figures.
As it can be seen our data is in an excellent agreement with the experiments. The simulation data obtained by Poprocki et.al does not agree with their experiments, while the Gras-Marti model only gives accurate recoil profile near the surface.After checking our simulation program wih the experimental data, we attempted to accurately represent the mixing interface by following the reversed Si recoils in the Au layer. That is, a complete profile of both recoil atoms is obtained. Figures 2 (a-c) are examples of these simulations. After running the program many times with different input parameters such as, ion species, energy and dose we chose to simulate the mixing in 20nm and lOnm inside the Si and Au layers respectively. We have divided the mixing layer into 0.5 nm sectors in depth and obtained the recoils in each layer. We suggest that the thickness of the mixing layer can be approximated from where the concentration of recoiled atoms is comparable to the density of the host layer, In this way, from figures (4-5) a mixed layer of maximum 5.0 nm is obtained.
W r have developed a program called TRIM-MIXING to simulate the intermixing of metal-silicon
bilayers. Recoil profiles obtained by this program is in an excellent agreement with the experiments. In addition, by following Si and Au atoms recoiled across the Au-Si interface we are able to define a more precise mixing layer.
Concentratioderb units)
'
*
TRW-MIXINGW Q-MT
Depth (nm)
Concenbation(arb units)
1 F 1
*
TRIM-MIXING 0.10.0 1
C.
0 A
I
0.001
0 20 40 60 80 100 120 140 Depth (nm)
Cancentratiodarb unite)
1 I
1 0 Experimental
A TRIM-MIXING
0.1 :
I
0.01 :
l
J I
Fig. 1
-
The profile of Au recoils in Si from experimental results of Poprocki et. a1 [l31 compared to Gras-Marti analytical data, computer simulation results by Poprocki et. a1 and data obtained using TRIM-MIXING for( 2 )
120 lcev Ar at a dose of 5xl0l'+ions\cm~,, (b) 2002kev Kr at a dose of 1014 ions/cm,
and (c) 300 kev Xe ata dose of 10 ions/cm
.
C$-284 COLLOQUE D E PHYSIQUE
Concentratlon(arb units)
.
1nk#8ce
Fig. 2
-
The profile of Au and Si recoils across Au/Si interface for (a) 120 kev Ar incident on 26nm-Au/Si bilayer, (b) 200 kev Kr incident on 29nm-Au/Si bilayer, and (c) 300 kev Xe incident on 25nm-Au/Si bilayer.References
/l/ Matteson, S., Paine, B.M., Grimaldi, M.G., Mezey, G. and M.-A. Nicolet, Nucl.Instr. and Meth. 182/183 (1981) 43.
/2/ Mayer, J.W., Tsaur, B.Y., Lan, S.S. and Hung, L.S., Nucl.Instr. and Meth. 182/183 (1981) 1.
/3/ Paine, B.M. and Averback, R.S., Nucl. Instr. and Meth. z / 8 (1985) 666.
/4/ Xia, W., Fermandez, M., Hewett, C.A., Lau, S.S., Poker, D.B. and Biersack, J.D., Nucl.
Instr. and Meth. B37/38 (1989) 408.
/5/ Hung, L.S., Hong, Q.Z. and Mayer, J.W., Nucl. Instr. and Meth. B37/38 (1989) 414.
/6/ Matteson, S., Paine, B.M. and Nicolet, M.A., Nucl. Instr. and Meth. W 1 8 3 (1981) 53.
/7/ Sigmund, P. and Gras-Marti, A., Nucl. Instr. and Meth. 182/183 (1981) 25. /8/ Sigmund P., Appl. Phys. A B (1983) 43.
/9/ Hung, L.S., Xia, W., Poker D.B., Fermandes, M., Tao, K., Lau, S.S and Mayer J.W., J. Appl.
Phys. 64 (1988) 2354.
/10/ Biersack, J.P. and Eckstein, W., J. Appl. Phys. A 34 (1984) 73.
/11/ Moller, W. and Eckstein, W., Nucl. Instr. and Meth. B2 (1984) 814.
/12/ Biersack, J.D. and Haggmark, L.G., Nucl. Instr. and Meth.174 (1980) 257.
/13/ Poprocki, K., Brylouska, J. and Syszko, W., Appl. Phys. A 45 (1988) 109.
/14/ Gras-Marti, A., Phys. State Solidi (a) 76 (1983) 621.
/15/ Abril, I., Garcia-Molina, R. and Gras-Marti, A., Phys. State Solidi (a)
