Track etching in amorphous metallic Fe81B13.5Si3.5C2

Nuclear Instruments and Methods in Physics Research B 107 (1996) 397-402

Beam Interactions with Materials 8 Atoms

ELSEVIER

Track etching in amorphous metallic Fe,,B && C. Trautmann

a** , C. Dufour ‘, E. Paumier

b*c,R.

Spohr

a,

M. Toulemonde

b

a Gesellschaf fir Schwerionenforschung mbH. Postfach 110552,64220 Darmstadt, Germany b CIRIL, laboratoire mixte du CEA et CNRS, Rue Claude Bloch, BP 5133,14040 Caen cedex, France ’ LERMAT, ISMRa, Blud du Mare&al Juin, 14032 Caen. France

Abstract A review of the presently available results for track etching in metallic glasses is given. Latent tracks in amorphous Fe,,B ,3,5Si3,5C2 can be selectively etched to surface pores with a wide cone opening angle. Etched tracks have been observed only if the alloy is in the amorphous and not in the crystalline state. Annealing tests after ion irradiation show that track fading occurs for temperatures above 200°C. An etching threshold is observed at a critical energy loss of 3.4 keV/A, which is far above the threshold for creating damage (1.3 keV/& as observed by resistivity measurements. The etching behaviour near the threshold shows a transition regime from non-homogeneous to homogeneous etching. This observation is related to etched tracks in insulators. The occurrence of the etching threshold is used to determine the influence of the distribution of the radial dose of ions having various velocities. The experimental results are compared with thermal spike calculations.

1. Introduction While ion tracks in insulators have been investigated for more than three decades [l-3], tracks in metals have only been known for a few years. Numerous experiments have provided clear evidence that the high electronic energy transfer (d E/dx) of energetic heavy ions induces radiation effects in various metals, whether the solid was in the amorphous [4-91 or in the crystalline state [lo-131. In different types of crystalline metals [lO,l l], a high rate of increase of the resistivity due to electronic energy loss has been observed. Imaging of single ion tracks by transmission electron microscopy (TEM) has only been successful for the crystalline metallic compound NiZr, [ 131 and for the pure metals Ti and Zr [ 141. In metallic glasses, modifications of the bulk material have been investigated by several techniques. One observed effect was anisotropic growth, whereby the irradiated samples shrank parallel to, and expanded perpendicular to, the direction of the ion beam if an incubation fluence was surpassed. Another ion-induced effect in amorphous metallic alloys is the increase of electrical resistivity at a high rate above a critical d E/dx threshold [8,9,23]. In both of these experiments, integral modifications of the bulk material due to a high particle fluence (above 10” ions/cm’) were studied. Small-angle neutron scattering (SANS) also allows the observation of many

’ Corresponding author. Tel. + 49 6159 7 127 16, fax + 49 6159 712179.

tracks indirectly, as has been demonstrated for insulators such as mica and polymers [ 151. However, tracks (Xe ions, 2.8 MeV/u) in the metallic glass Pd,,Si,, could not be observed using this technique [ 161. Due to the non-existence of a regular lattice as in a crystal, TEM is not suitable for the observation of single ion tracks in amorphous solids. In such materials the only available technique for studying individual tracks is to enlarge them by chemical etching, as has been performed, for example, in vitreous SiO, [17]. This process dissolves the damaged material along the ion track at a higher rate than the non-irradiated bulk material. Here, we report the numerous results obtained by etching tracks of various ions in the glassy iron-boron alloy Fe,,B,,,sSi,,,C,. In addition, annealing experiments were performed before and after ion irradiation combined with etching.

2. Irradiation For all experiments we used 25-30 &rn thick foils of amorphous metallic Fe s, B ,3,5Si3,5C2 supplied by Goodfellow (181. This alloy was prepared by melt spinning with a quenching rate of lo6 K/s. The irradiations were performed for medium (Xe, Dy, Au and U ions) and high (Au and Bi ions) energies at the linear accelerator Unilac and at the heavy ion synchrotron SIS at GSI (Darmstadtl. Some of the U ion irradiations were performed at GANIL (Caen). The energy of the ions was in all cases high enough that the ions passed through

0168-583X/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0 168-583X(95)00848-9

V. RELATED

FIELDS

398

C. Trautmann et al./Nucl.

Instr. and Meth. in Phys. Res. B 107 (1996) 397-402

the foils. At Unilac, the ion energy was varied by using polymer foils as absorbers, which were etched afterwards in order to control the applied ion fluence. At SIS energies, stacks of up to 15 foils of Fe,,B,,,Si,,C, were irradiated. With the computer code TRIM [ 191, using a density of 7.32 g/cm3 for this alloy [18], the energy, the stopping power, and the residual range of the ions were determined on the front side of each foil. During all irradiations a hexagonal structured grid was placed in front of the samples. Since the ions were stopped in the grid, this resulted in a characteristic pattern of irradiated and non-irradiated regions. The applied ion fluence was in the range lo6 to 10’ ions/cm2. All irradiations took place at ambient temperature.

3. Track etching Successful track etching requires that the etching rate ut along the track is greater than the etching rate ub of the non-irradiated bulk material. The ratio of the two etching rates ut,/ut determines the geometry of the etched pores. In general, the bulk etching rate remains constant for a given material and for a given etchant applied under a specific set of etching conditions such as concentration and temperature. In addition to the factors mentioned, the track etching rate depends on the density of damage in the track and hence on the energy and on the energy loss of the track-forming ion. Tracks in amorphous metallic Fe,,B,,,Si,,C, were selectively etched in a 0.04N FeCl, solution [20]. More recently, better results have been obtained in a 0.7N HCl solution while applying a DC voltage of about 0.8 V. Although the surface of this alloy was quite rough, successful track etching was recognised when, due to the significant pattern of the mask, irradiated and non-irradiated areas could clearly be distinguished. Single etched pits

Fig. 1. SEM image of single tracks of Bi ions (E = 28.8 MeV/u, dE/dx = 4.3 keV/A, fluence = lo6 ions/cm*) in Fe,,B,3., Si 3.5C 2 after etching in a 0.7N HCI solution (5 min, applied voltage - 0.8 VI.

Fig. 2. SEM image of a Fe,,B ,3.5Si&r sample (irradiated with IO’ Au ions/cm*, E = 11.6 MeV/u, dE/dx = 5.2 keV/A) an-

nealed for 1 h at 200°C prior to etching. Irradiated (top right and left) and masked areas can clearly be distinguished.

were studied by scanning electron microscopy @EM) (Fig. 1). For almost all samples, the number of pores revealed corresponded to the applied fluence. In contrast to polymers, where an extremely small vr,/vt ratio leads to almost cylindrical pore geometry, etched tracks in the investigated metallic alloy gave conical pits with a wide opening angle of around 60”. This geometry corresponded to an etching ratio of VJU, = 1:2. It is interesting to note that a similar pore geometry has been obtained for etched tracks in amorphous soda lime glasses [I]. 3.1. Annealing

Several annealing experiments were performed prior to irradiation. To test whether tracks in Fe,,B,,,Si,,sC, were etchable if the alloy was in the crystalline state, the metal was annealed for 140 min at a temperature of SlO”C, which is well above the crystallisation temperature (r, = 480°C) of this alloy. With such treatment, crystallisation took place [21], the foils lost their ductility and became very brittle. After irradiation with U ions (11.4 MeV/u, d E/d x = 6.3 keV/A> we could not find any etched tracks. Other samples were annealed for 720 min at a temperature of 200°C without changing the glassy state. After this treatment etched pores showed no obvious difference in pore shape or size compared with the untreated samples. In a third experiment, the influence of temperature on track stability was studied. Several samples irradiated with Au ions (11.4 MeV/u, dE/dx = 5.2 keV/A) were kept for 1 h at an increased temperature (50°C 5 TS 4OO”C), followed by a slow cooling procedure at a rate of about 30 K/h. In the temperature range between 50 and 200°C the pattern of the irradiation mask was visible, but became more and more diffuse with increasing temperature. 200°C was the highest temperature where irradiated and nonirradiated zones could be distinguished (Fig. 2). After

C. Traunann et al./Nucl. Instr. and Meth. in Phys. Res. B IO7 (1996) 397-402

399

3.6 keV/A, the number of pores revealed only corresponded to 70% of the ion fluence applied. The mean diameter of etched pores was smaller and had a wider distribution than pores at higher energy losses. Above an energy loss of d E/d x 2 3.6 keV/.&, etched pores were of uniform size. At slightly increased values of the d E/d x threshold, similar behaviour was observed in the case of Au and Bi ions (Fig. 3). 0

10

20

30

40

50

60

70

80

4. Discussion

specific energy oMev/u) Fig. 3. Calculated curves [19] of the energy loss for Xe, Dy, Au, Bi and U ions m Fe,,B,,,, Si r,s C r as a function of their specific energy. The data correspond to etching experiments: only in the case of filled symbols was homogeneous etching possible. In the transition region (EB) non-homogeneous etching was observed and at lower energy losses (open symbols) tracks could not be etched.

applying

of 230°C

a temperature

or above,

etched

The observation that tracks have been etched only as long as the samples were in the amorphous phase strongly suggests that the phase of the metal plays an important role. From the etching experiments it could not be concluded whether the tracks were formed or not, or whether the damage density was not sufficient to be etchable. In resistivity experiments, it has been observed that metallic alloys are less sensitive when they are in the crystalline rather than in the amorphous state [23]. For crystalline iron, an extremely high threshold (d E/d x > 4 keV/& for creating damage due to electronic energy loss has been found [24]. It is not clear whether an homogeneous damage density, as required for etching, can be achieved for crystalline iron, even if the highest possible d E/d x of U ions is used.

tracks

could not be identified. 3.2. Injluence of energy loss In many solids, a significant property of track etching is the existence of an etching threshold. It is defined as the minimum energy loss that is required to form an etchable track. In amorphous FesiB,s,,Si,,.$,, etched tracks were revealed under typical etching conditions only in those cases where the electronic energy loss surpassed a level of 3.4 keV/A [22]. Fig. 3 gives theoretical curves for the energy loss [ 191 of various ions as a function of their specific energy superimposed with the experimental data. For Xe ions, ~0 pores were found for tn energy loss below 3.4 keV/A. In the range 3.4 keV/A < d E/dx <

8 , g

I

I

/

I 6-

4.1. Etching threshold and track morphology The phenomenon of non-homogeneous etching has been observed previously in insulators such as SiO, [ 171, polymers [25] and yttrium-iron garnet (Y3Fe,0,,) [26] and has indicated that the concentration of the defects was non-homogeneous, but varied along each track. Especially for

/

/

I

y3Fe50,,

.’

1

0

Y . . ,’ . 6 “a-Fe,,B,, 5 5 _ c-SiO, . . ,’ _ 6 ti 40 8 :~:_~------i~ ; 3--o-I d< ‘3 W 2:_____,L________:______

track morphology

n

$ 0

N

I

q lmoO - ~O,,or_

0

0

B I 1

,O’ -

I 2

-

-

-

_

_

_

I I 1 3 4 5 energy loss (keV/A)

_

_

, 6

_

_.

_

III

_

_

IIBO 0

_

_

0

0.6 nm

7

Fig. 4. Effective track radius increases as a function of the energy loss for insulators (circles, SiO, quartz 117,291; squares, Y,Fe,O,, [26-281) and for the amorphous metal Fe,,B,, (diamond) [8,30]. Track etching is only observed for tilled symbols. The dashed line corresponds to a fit to the resistivity data of amorphous Fe,,B,,. The threshold (A) for homogeneous etching of Xe ions in It can be correlated to a track radius of 3.5 nm. The scheme on the left shows the Fes,Bis., Si 3.5 C z of this work is superimposed. morphology of the damage zone along the ion track according to studies of tracks in Y,Fe,O,, [28].

V. RELATED

FIELDS

ml

C. Trautmann et al./Nucl.

Instr. and Meth. in Phys. Res. B 107 (1996) 397-402

yttrium garnet, the morphology of the damage zone, as a function of the energy loss, is well known due to detailed studies with various techniques [26-281. On the left side of Fig. 4, the morphology of the defects along the ion track is shown for this material. Four distinct regions are observed corresponding to different effective track radii. For small radii at low energy losses (regime II), the defects were spherical and well separated. With increasing energy loss, they percolated forming extended defects (regime III) and became cylindrical, but were still discontinuous. In regime IV, the extended defects became continuous, whereby the concentration of the damage was not yet homogeneous. When track etching became possible, however, the diameter of the pores showed a wide distribution in size [26]. For higher energy losses (regime V), the defects were continuous cylinders with a homogeneous density of damage. In this regime the etched pore size was uniform. Following the morphological description of latent tracks in yttrium-iron garnet, the observed threshold (d E/d x = 3.6 keV/W> for homogeneous track etching in Fe,,B,,,sSi,,,C, could be ascribed to the forming of a continuous cylindrical damage zone alone the ion track, while the lower limit (d E/d x = 3.4 keV/A) corresponded to a non-homogeneous damage zone. Below the threshold the defects were more or less discrete and could not be revealed by etching. Since the etching procedure destroys the latent track, no direct information about the track radius could be obtained. However, there is data available, which has been obtained from resistivity measurements in amorphous Fe,sB,S [S]. In this alloy, the damage cross section S, leads to an effective track radius R, (with Rzn = S,, Fig. 15 of Ref. [30]), which is plotted versus the energy loss in Fig. 4. Accorting to those results, the observed threshold of 3.6 keV/A for homogeneous etching of Xe tracks corresponded to a track radius of about 3.5 nm. Hereby it was assumed that the resistivity data for Fe,,B,, could be applied to the slightly different alloy Fes,B,,,,Si&, used in this experiment since various amorphous metallic alloys, containing at least 80% iron, showed the same dimensional changes under ion irradiation 171. 4.2. Etching threshold and ion velocity In insulators it has been clearly shown that not only the energy loss, but also the radial distribution of deposited energy, has to be taken into account [28,31]. For a given value of energy loss, the energy deposited on the electrons spreads out to a larger distance at higher ion energies [32,33]. Experiments with ions of various velocities were performed to quantify this effect on the threshold of track etching for metallic Fe,,B,,,Si,,C,. For track etching, a critical damage density had to be reached corresponding to the etching threshold. At a higher ion energy it was expected that the critical damage density would be reached at a higher energy loss. Fig. 5 shows the threshold of

I

0

I

I

I

I

I

I

I

I

I

I

I

I

10

20

30

40

50

60

70

energy (MeV/u)

Fig. 5. The experimental etching thresholdfor Xe, Au and Bi ions as a function of energy compared with calculations (solid line) using the thermal spike model.

homogeneous track etching for Xe, Au and Bi ions versus the energy at which this threshold was obtained. The influence of the radial energy distribution is not very pronounced. The observed threshold slightly increases as a function of the ion energy. Due to the inaccuracy of the experimental data, which mainly originated from the d E/d x error, the etching threshold as a function of the ion energy was also compatible with a constant. These experimental observations were compared with calculations performed with the thermal spike model [34,35]. Assuming the track corresponds to a cylinder of a molten phase, it is possible to calculate the track radius as a function of the ion energy loss using the electron-phonon coupling constant as a free parameter. A detailed description of the thermal spike calculations is given in Refs. [36,37]. For such calculations, we used the thermodynamic parameters of amorphous FessB,, [36] and the coupling constant obtained by fitting the track radii deduced from resistivity data, as presented in an earlier paper [34]. The threshold needed for homogeneous track etching could then be attributed to a critical track radius. In the next step, the etching threshold d E/dx (corresponding to this fixed critical track radius) was calculated as a function of the ion energy. For the initial spatial energy distribution of various ion energies, the analytical formula developed by Waligorski et al. was applied [33]. Fig. 5 shows that the experimental data for Xe, Au and Bi ions are in agreement with the evolution of the etching threshold as predicted by thermal spike calculations (solid line). At present, more detailed conclusions are not possible due to the lack of precision of d E/d x.

5. Conclusions A threshold for chemical etching of tracks in Fe,,B,,,Si,,&, was observed at an extremely high value of dE/dx of - 3.4 keV/A. This is the highest value ever

C. Trautmann et al./Nucl.

Instr. and Meth. in Phys. Res. B 107 (1996) 397-402

observed for any etching threshold and is possibly the clue to why etched tracks in metals have not been found until now. According to FessB ,s data [8], the latent track radius required for chemical etching is about 3.5 nm. It is interesting to note that the same track size was needed for chemical etching in the case of other solids such as Y,Fe,O,, or SiO,. After irradiation, track etching was inhibited for samples annealed at temperatures above 200°C. Annealing prior to irradiation did not affect the process if the temperature was kept below the crystallisation temperature. Contrary to the alloy in the amorphous phase, the same sample

in the crystalline phase did not show any etched tracks. This clearly indicates that the same material in its crystalline phase was less sensitive to the energy loss of the ions than if it was in the amorphous phase. This finding is in agreement with results obtained by resistivity measurements in Ni,B [23]. Tracks in Fe,,B,,,sSi,,,C,, revealed near the threshold, showed a larger inhomogeneity compared with tracks produced at higher d E/d x. This behaviour was observed for various insulators (e.g., Y3Fe,0,, [26], SiO, [17], polymers [25]). It is in agreement with the model of discontinuously extended defects, which was developed to describe the track morphology in insulators. Below the threshold, the defects along the ion tracks were more or less discrete and could not be revealed by etching. The defect morphology explains the fact that higher energy losses were needed for chemical etching than were observed for creating dam-

age as seen by resistivity measurements. The occurrence of the etching threshold

was used to test the effect of the radial distribution of deposited energy. This was done by comparing the thresholds for homogeneous etching of Xe, Au and Bi ions at an energy of 6.1, 28 and 35.8 MeV/u, respectively. A slight increase of the etching threshold was observed, but the present knowledge of the accuracy of the energy loss values was not precise enough to confirm quantitatively the increase as obtained by thermal spike calculations.

Acknowledgements We would like to thank K. Dermati for performing annealing experiments.

the

References

111R.L. Pleischer, P.B. Price and R.M. Walker, Nuclear Tracks in Solids, Principles and Applications (University of California Press, Berkeley, 1975). La S.A. Dutrani and R.K. Bull, Solid State Nuclear Track Detection (Pergamon, Oxford, 1987). 131R. Spohr, Ion Tracks and Microtechnology. Basic Principles and Applications (Vieweg, Braunschweig, 1990).

401

[41 S. Klaumlinzer,

Ming-Mong Hou, G. Schumacher and Chang-Lin Li, Mater. Res. Sot. Symp. Proc. 93 (1987) 21. 151S. Klaumhnzer, G. Schumacher, S. Rentzsch, G. Vogl, L. Sijldner and H. Beiger, Acta. Metall. 30 (1982) 1493. 161S. Klaumiinzer, Ming-dong Hou and G. Schumacher, Phys. Rev. Lett. 57 (1986) 850. 171Ming-dong Hou, S. Klaumilnzer and G. Schumacher, Phys. Rev. B 41 (1990) 1144. h31A. Audouard, E. Balanzat, J.C. Jousset, D. Lesueur and L. ‘Ihome, J. Phys. Condensed Matter 5 (1993) 995. [91 A. Audouard, E. Balanzat. G. Fuchs, J.C. Jousset, D. Lesueur and L. Ihome, Europhys. Lett. 5 (1988) 241: Europhys. Lett. 3 ( 1987) 327. 1101A. Dunlop and D. Lesueur, Radiat. Eff. and Def. in Solids 126 (1993) 123. [ill C. Dufour, A. Audouard, F. Beuneu, J. Dural, Y. Girard, A. Hairie, M. Levalois, E. Paumier and M. Toulemonde, J. Phys. Condensed Matter 5 (1993) 4573. 121 A. Audouard, E. Balanzat, S. Bouffard, J.C. Jousset, A. Chambered, A. Dunlop, D. Lesueur, G. Fuchs, R. Spohr, L. Thorn& and J. Vetter, Nucl. Instr. and Meth. B 59/60 (1991) 414. 131 A. Barbu, A. Dunlop, D. Lesueur and R.S. Averback, Europhys. Lett. 15 (1991) 37. 141 H. Dammak, A. Dunlop, D. Lesueur, A. Brunelle, S. DellaNegra and Y. Le Beyec, Phys. Rev. Lett. 74 (1995) 1135. 1151D. Albrecht, P. Armbruster, R. Spohr, M. Roth, K. Schaupert and H. Stuhrmamt, Appl. Phys. A 37 (1985) 37. S. Klaumiinzer, W. Petty and G. Schumacher, Nucl. Tracks [I61 Radiat. Meas. 19 (1991) 907. [171 A. Sigrist and R. Balzer, Helv. Phys. Acta 50 (1977) 49. Goodfellow GmbH, Kursstr. 13-15, 61231 Bad Nauheim, 1181 Germany. [191 J.F. Ziegler, J.P. Biersack and U. Littmark, The Stopping and Ranges of Ions in Solids (Pergamon, New York, 1985). DOI C. Trautmamr, S. Andler, W. Brlichle. R. Spohr and M. Toulemonde, Radiat. Eff. and Def. in Solids 126 ( 1993) 207. J.L. Walter and H. El1 W.G. Johnston, A. Mogro-Campero, Bakhru, Mater. Sci. Eng. 55 (1982) 121. [=I C. Trautmann, R. Spohr and M. Toulemonde, Nucl. Instr. and Meth. B 83 (1993) 513. [231 A. Audouard, E. Balanzat, S. Bouffard, J.C. Jousset, A. Chamberod, A. Dunlop, D. Lesueur, G. Fuchs, R. Spohr and J. Vetter, Phys. Rev. Lett. 65 (1990) 875. [241A. Dunlop, D. Lesueur, P. Legrand and H. Dammak, Nucl. Instr. and Meth. B 90 (1995) 330. [=I C. Trautmamr, IRAP 94 Conf. Proc., Nucl. Instr. and Meth. B 105 (1995) 81. [261M. Toulemonde, N. Enault, Jim Yun Fan and F. Studer, J. Appl. Phys. 4 (1990) 68. [271 F. Studer, C. Houpert, H. Pascard, R. Spohr, J. Vetter, Jin Yun Fan and M. Toulemonde, Radiat. Eff. and Def. in Solids 116 (1991) 59. Dl A. Meftah, F. Brisard, J.M. Costantini, M. Hage-Ali, J.P. Stoquert, F. Studer and M. Toulemonde, Phys. Rev. B 48 (1993) 920. [291M. Toulemonde, E. Balanzat, S. Bouffard, J.J. Grob, M. Hage-Ah and J.P. Stoquert, Nucl. Instr. and Meth. B 45 (1990) 64. [301 A. Audouard, E. Balanzat, J.C. Jousset, G. Fuchs, D. Lesueur and L. Thorn&, Mater. Sci. Forum 97-99 (1992) 631.

V. RELATED

FIELDS

402

C. Trautmann et al./Nucl.

Instr. and Meth. in Phys. Res. B 107 (1996) 397-402

[31] A. MetIah, N. Merrien, N. Nguyen, F. Studer, H. Pascard

and M. Toulemonde, Nucl. Instr. and Meth. B 59 (1991) 605. [32] R. Katz and E.J. Kobetich, Phys. Rev. 186 (1969) 344. [33] M.P.R. Waligorski, R.N. Hamm and R. Katz, Nucl. Tracks Radiat. Meas. 11(1986) 309. [34] C. Dufour, Thtse de J’Univetsid de Caen (1993).

[35] C. Dufour, E. Paumier and M. Toulemonde, Radiat. Eff. and Def. in Solids 126 (1993) 119. [36] M. Toulemonde, C. Dufour and E. Paumier, Phys. Rev. B 46 (1992) 14362. [37] C. Trautmann, C. Dufour, E. Paumier and M. Toulemonde, to be published in Nucl. Instr. and Meth.