- Email: [email protected]
Application of mass-separated focused ion beams in nano-technology
Available online at www.sciencedirect.com
NIM B Beam Interactions with Materials & Atoms
Nuclear Instruments and Methods in Physics Research B 266 (2008) 1846–1851 www.elsevier.com/locate/nimb
Application of mass-separated focused ion beams in nano-technology L. Bischoff * Research Center Dresden-Rossendorf Inc., Institute of Ion Beam Physics and Materials Research, P.O. Box 51 01 19, D-01314 Dresden, Germany Received 15 November 2007 Available online 15 December 2007
Abstract FIB applications like writing ion implantation, ion beam mixing or ion beam synthesis in the lm- or nm range often require ion species other than gallium. Therefore alloy liquid metal ion sources (LMIS) have to be developed and applied in FIB tools. The energy distribution of ions emitted from an alloy LMIS is one of the crucial parameters for the performance of a FIB column. Different source materials like AuGe, AuSi, AuGeSi, CoNd, ErNi, ErFeNiCr, MnGe, GaBi, GaBiLi, SnPb, . . . were investigated with respect to the energy spread of the different ion species as a function of emission current, ion mass and emitter temperature. Different alloy LMIS’s have been developed and used in the FZD – FIB system especially for writing implantation to fabricate sub-lm pattern without any lithographic steps. Co and various other ion species were applied to generate CoSi2 nano-structures, like dots and wires by ion beam synthesis or to manipulate the properties of magnetic films. Additionally, the possibility of varying the flux in the FIB by changing the pixel dwell-time can be used for the investigation of the radiation damage and dynamic annealing in Si, Ge and SiC at elevated implantation temperatures. Furthermore, a broad spectrum of ions was employed to study in a fast manner the sputtering process depending on temperature, angle of incidence and ion mass on a couple of target materials. These studies are important for the 3D-fabrication of various kinds of micro-tools by FIB milling. Ó 2008 Elsevier B.V. All rights reserved. PACS: 68.03; 61.72; 85.40Ry; 85.45Bz; 71.55i Keywords: Alloy liquid metal ion sources; Energy spread; Focused ion beam; Application; Microstructures
1. Introduction With the invention of the liquid metal ion source (LMIS) in the sixties the focused ion beam (FIB) technique started an impressive development from the laboratory level to high performance industrial equipments [1,2]. Applying these unique point-like sources with a brightness in the order of 106 Acm2 sr1 ion beams with a diameter less than 10 nm and current densities more than 10 Acm2 can be achieved. So a very promising field was opened to use the FIB for mask [3] and integrated circuit repair and modification [4], failure analysis [5] or TEM specimen preparation [6], as well as in material science for sputtering
*
Tel.: +49 351 260 2963; fax: +49 351 260 3285. E-mail address: l.bischoff@fzd.de
0168-583X/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2007.12.008
investigations and applications [7,8], for the formation of quantum dots and wires [9] for the fabrication of MicroElectro-Mechanical Systems (MEMS) [10] and for a lot of other applications in micro- and nano-technology. The majority of FIB systems operating worldwide are suited only for gallium ions, due to the favoured properties, like its low melting point (Tm = 29.6 °C) and the simple emitted mass spectrum. About 99% of the ions in the beam are singly charged. Due to the mass-independent behaviour of electro-static ion optics, no mass filter is needed. The two isotopes have not to be separated and the ion column design becomes easier. But there is an increasing demand for ion species other than Ga. Especially, in order to use the advantages of the FIB, in terms of the high lateral resolution, the flexibility in fluence and pattern design for direct doping purposes, one needs different ion sources to produce the wealth of technologically interesting ions. As
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851
an example B+, As+ or P+ for Si-doping [11] or rare earth elements like Er+, Pr+ and Nd+ for opto-electronic applications [12] metals, like Au+ for plasmonic structures [13] or Fe+, Co+ and Ni+ for selective changes of magnetic properties [14] should be mentioned. To operate LMISs with such source materials is in most cases difficult, due to the high melting points of the pure materials. The composition of suitable binary or ternary alloys and a compatible emitter material can overcome this problem. It should be mentioned, that the application of alloy LMISs requires also an ion optical column equipped with a mass separator to select the wanted ionic species as well as to separate the certain isotopes. The development and preparation, as well as the characterisation of alloy LMISs, and examples of the applications of mass-separated focused ion beams will be summarised in this article. 2. Alloy liquid metal ion sources To focus an ion beam into a spot size of a diameter smaller than one lm a source is needed which emits the ions from a very small area into a limited solid angle. LMISs as well as gaseous field ionisation sources approach these conditions [15]. Because of their broad and versatile applications the following discussion is concentrated on the LMIS only. This type of sources usually consists of an emitter needle with a tip radius of about 5–10 lm, which is covered with a pure metal, i.e. Ga, In, . . . or an alloy, i.e. Au82Si18, Co64Nd36, . . . and heated directly or indirectly. A counter electrode (the extractor) is placed in front of the needle where a high voltage (extraction voltage) in the range of 2–10 kV is applied. At a critical field strength of about some 10 V/nm on the tip a liquid cone, the so called Taylor-cone [16] with a half angle of 49.3° is formed and ion emission occurs from a liquid jet on this cone, mainly due to a field evaporation process [17]. In the following we want to deal more in detail with alloy LMIS. To find a suited material for the desired purpose one have to look for a low temperature melting alloy (mostly in the eutectic composition) with a low vapour pressure and chemical inertness with the source base material. The source preparation itself may be divided in three basic steps: the fabrication of the emitter, the electro-chemical etching using a NaOH solution in the case of tungsten or a mechanical treatment for tantalum emitters to form the tip and the wetting and reservoir filling of the emitter by dipping in a crucible containing the molten alloy in a high vacuum ambient. Table 1 gives an overview about the investigated alloy sources, their melting points and the main application fields. It should be mentioned that all materials exhibit a vapour pressure lower than 108 mbar at the operation temperature beside manganese. A description of the preparation procedure can be found more in detail elsewhere [2,12,18].
1847
Table 1 Alloy composition, melting temperature and application of investigated (alloy) LMIS in the FZD – FIB Material
Melting temperature (°C)
Applications
Au73Ge27
365
Au77Ge14Si9
365
Au82Si18 Co36Nd64
365 566
Co27Ge73 Er69Ni31 Er70Fe22Ni5Cr3 Sn74Pb26 In14Ga86
817 765 862 183 14.2
Mn45Ge54Si1
720
Ga38Bi62 Ga35Bi60Li5
222 250
Au: sputtering, nano-cluster; Ge: Doping, SixGe1x alloy Si: contamination free processing, imaging dto Co: CoSi2 – IBS, Nd: optical + magnetic application Co: magnetic application Er: optical application Fe,Ni,Cr: magnetic application Sn: doping, fundamental research Fundamental source emission investigation Mn implantation in compound semiconductors Ga: acceptor; Bi: shallow donor Li: source for analysis and ion beam lithography All classical FIB applications
Ga
29.7
The primary investigation is the recording of the I-Vcharacteristics from which the onset voltage of the source and its stability can be obtained. Additional the slope of this curve is important for the control adjustment of a stable emission. Analysing the I-V-characteristics as a function of temperature yields the optimum operation point reflecting the behaviour of the surface tension of the melt [19,20]. Normally for emitters, wetted with AuGe or CoNd the extraction voltage decreases with temperature, according to theoretical predictions [21], due to a corresponding decrease in the surface tension coefficient of the alloy. In the case of silicon containing emitters (AuSi, AuGeSi) an entirely different behaviour was found. This behaviour is, in turn, associated with the anomalous behaviour of the surface tension coefficient, arising from a residual crystalline structure at the liquid surface [22]. Therefore, the extraction voltage increases with temperature up to about 800 °C and then decreases as expected slowly at higher temperatures. The energy spread DE of the different ion species is a very important parameter which is significant in determining the final resolution of a FIB arising, in turn, from the chromatic aberrations [2,21]. This parameter was analysed as a function of emission current I, ion mass m and emitter temperature T in the current range of 1–30 lA. For singly charged ions the predicted dependence of the energy spread, DE / I2/3 m1/3 T1/2, found for elementary monoatomic LMIS [24,25] could be reasonable confirmed. For sources, emitting a multiplicity of ion species with different charge states, molecule ions and clusters other dependencies may be obtained [26]. For example, a singly charged Co beam from a CoNd alloy LMIS shows a dependence
1848
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851
Energy spread for differnt ion species I emission = 15 µA
+
Au5
100
+
Au4 +
ΔE FWHM (eV)
Au3 +
Au2
++
Au ++ Er Ni ++ ++ Nd + ++ Co ++Sn + Au ++ Ge Sn Fe + + Ni Ga + Co + + Si Fe ++
++
Si
10
AuGeSi + AuGe
x
fit ΔE ~ m single charged x = 0.33 double charged x = 0.21 cluster ions x = 0.86
+
Li
10
+
100
1000
m/q (amu/e) Fig. 1. Energy spread of singly-charged ions, clusters, and doubly-charged ions as a function of their mass-to-charge ratio obtained from alloy LMIS AuGeSi, AuGe, ErFeNiCr, CoNd, GaBiLi and from pure metal LMIS for Sn and Ga at a constant emission current of 15 lA [20]. The accuracy of the measurements is about 1 eV.
of DE / I0.4 which is in good agreement with some experimental data from [23] and predictions of Mair [26] but in contradiction to the calculations of [24,25]. But the first could be confirmed in the case of an AuSi alloy LMIS. The dependence of DE on the ion mass is presented in Fig. 1. For singly charged ions, DE is proportional to m1/3 which is in good agreement with the theory of Knauer [24] and Kim et al. [25] as well as with the experimental results of Swanson [27]. For doubly charged ions a weaker mass-dependence was found but the absolute values are higher than that of singly charged ones. This behaviour was also found by Marriott [28] and Ishitani
1000
Bi
et al. [29] who indicated the general relationship DE(M2+)/2 < DE(M+) < DE(M2+). For clusters a stronger slope was determined but the data in the literature are rather sparse for a safe comparison. Another interesting point is the composition of the species coming from the alloy wetted emitter. This can be determined by analysing the mass spectrum. A typical high resolution spectrum from an GaBiLi alloy LMIS is shown in Fig. 2 [30]. As can be seen, the single charged ions Bi+ and Ga+ are the dominant species in the beam. For an ion of charge state n, Brandons criterion [31] state that the evaporation field E determines which ion species will
Ga33Bi57Li10 -LMAIS
+
+
Ga
7
Li
+
6
Bi
++
+
Bi4
1
GaBi ++
10
GaBi +
current (nA)
+
Bi 3 Bi 2 Bi 3++
100
+
++
Ga
0.1
50
100
7
+
Li2
200
ExB voltage (V) Fig. 2. Mass spectrum of a Ga45Bi45Li10 alloy LMIS [30].
+
Li
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851
dominate in the beam. If E(I2+) < E(I+), then the surface atom is likely to be field evaporated as a doubly charged ion. But in our case E(I2+) > E(I+) so the single charged ions dominate in the spectrum. 3. Applications of mass-separated FIB The experimental applications, presented here, were performed using partly the in house developed mass-separated FIB system IMSA-100, which is described in detail elsewhere [32], and the modern high resolution FIB equipped with an ion-optical column CANION 31 Mplus (Orsay Physics). As a first example the FIB Ion Beam Synthesis of CoSi2 nanostructures will be described. Buried, single-crystalline CoSi2 layers were produced first by White et al. using ion beam synthesis (IBS), i.e. high fluence Co implantation into a heated silicon substrate followed by a two step annealing [33]. The crucial point was the implantation at elevated target temperatures to avoid amorphisation of the silicon substrate in order to submit the crystalline orientation information to the epitaxial growing silicide. A couple of years ago this process was transferred to the writing FIB technology to obtain smaller structure dimensions [34]. Therefore a Co36Nd64 alloy LMIS was developed [35] and low resistivity, single-crystalline CoSi2-microstructures were successfully fabricated down to dimensions as small as 60 nm [36]. The application of the high resolution FIB admitting a Co2+ beam (60 keV) with a spot size <50 nm opened the possibility to fabricate nanowires (NW) with a width down to 20 nm [37]. Fig. 3 shows a SEM image of such a NW after removing the top Si layer by RIE. But with decreasing NW dimension, the NW growth becomes instable and the NW do not follow the FIB trace in every case. It is always adjusted along the h1 1 0i direction according to energetic reasons. To investigate this phenomenon more in detail lines were written with different ion species to create damage traces. Then Co was diffused into the wafer from the rare side [38] and the resulting NWs looked similar to those in the case of Co FIB implantation. That means, the NW growth is mainly driven by a damage related process, which is more precisely
<110> 20 nm
Fig. 3. SEM image of a 20 nm CoSi2 NW on silicon. The Co implantation fluence was 2.5 1016 cm2 [37].
1849
7 Nd- isotopes not separated 142 143 144 145 146 148 150
27.2 % 12.2 % 23.8 % 8.3 % 17.2 % 5.7 % 5.6 %
fluence variation
Σ = 1x1017 cm-2
Fig. 4. SEM image of CoSi2 NW and nanostructures in silicon after irradiation, Co-diffusion and annealing. Used was a selected Nd2+-FIB where all seven isotopes are separated but simultaneous present on the target after. The over all fluence was 1 1017 cm2.
explained in [39]. Fig. 4 demonstrates a FIB irradiation with a weak adjusted ExB filter using a 60 keV neodymium line resulting in seven small parallel lines corresponding to the natural abundance of the isotopes with an integral fluence of 1 1017 cm2. Depending on fluence and the distance to the neighbour line different shapes and lengths of CoSi2 NWs and nanostructures were obtained. Another problem of the FIB implantation in the nanoscale compared to the broad beam technique (BB) is the current density which can achieve 5–7 orders of magnitude higher values than that of BB implanters and exceeds 10 A cm2 corresponding to a flux of about 1019 ions cm2 s1 in the standing beam. If the spot is very fast moving over a sample surface one gets an averaged flux of the same order as a BB of about 1012 ions cm2 s1. Thus, the crucial parameter adjusting the flux is the pixel dwell time (PDT). This was used for extended investigations of the influence of these extreme flux to the dynamic annealing process after a channelling FIB implantation into Si and SiC [40]. For these experiments a Ge-beam was employed in order to avoid a sample doping or contamination. Varying the PDT (flux) and the target temperature and studying the shape of the implanted profiles by SIMS one can estimate the time scale for the relaxation of complex defects [41]. In both materials, at temperatures of about 200–250 °C significant defect reduction was found within the first 10 ls as well as between 10 ls and 100 s after ion impact. Whereas, at these temperatures the complex defects disappear nearly completely in Si, in SiC some of them survive. Similar investigations were carried out on germanium [42]. A well known and broad field of FIB application is the direct patterning by ion milling also called sputtering [43]. The escape of atoms, ions or clusters from the surface arising from a physical knock-on process caused by an incident ion beam depends on the ion mass and energy as well as the properties of the target material, like structure, density or surface binding energy, and also on the process parameters of temperature and angle of incidence. A theoretical
1850
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851 6
Sputtering yield depending on the ion mass
197
Au
target: Si, incidence angle: 0˚, energy: 30 keV
5
144
Nd
sputtering yield
4
74
3
69 59
Co
Ga
Ge
2
calculated experimental data
28
Si
1
0
0
20
40
60
80
100
120
140
160
180
200
220
ion mass (amu) Fig. 5. Calculated (according to Sigmund [44]) and experimental data of the sputtering yields as a function on ion mass for a 30 keV FIB with single charged ions in silicon for normal incidence.
description of this phenomenon can be found elsewhere [44,45]. The FIB is a very suitable tool for sputtering [1,20] and the characteristic parameter is the sputtering yield. Fig. 5 shows the calculated [44] and experimental data of sputter yields as a function on ion mass for a 30 keV FIB and single charged ions. It is obvious that switching from the usual Ga FIB to gold ions the sputtering efficiency increases by a factor of two which is suitable for the fabrication of micro-tools [46]. On the other hand turning to silicon or lighter ions only weak sputter erosion occurs at normal incidence which is adventurous for low damage imaging and ion lithography applications. In the last section of this review some other selected additional activities using a mass-separated focused ion beam from alloy LMIS will be mentioned to show the capability of this method and also to inspire new applications. Magnetic properties of very small dimensions are of increasing interest for high density storage media. It was observed that in Fe–Cr films the metastable paramagnetic phase can be transformed into a more stable b.c.c. phase which is ferromagnetic. Such a ferromagnetic pattern was written by a fine-focused Cr–ion beam in a paramagnetic alloy (face-centred orthorhombic, 37 at.% Cr) using an alloy liquid metal ion source (Er70Fe22Ni5Cr3). The written pattern were detected by magnetic force microscopy [47]. Furthermore a 60 keV Co2+-FIB was used to change the coupling on a lm-scale in a Ni81Fe19/Ru/Co90Fe10 film from antiferromagnetic to ferromagnetic behaviour [14]. Higher melting point alloys than Ga (i.e. Au73Ge27 and Au77Ge14Si9) can be solidified by quenching. Doing this during operation of an alloy LMIS at which the protrusion on the needle can be frozen a solid nm-tip is formed, which
can be used as an electron emitter after changing the polarity. To regenerate the source this process can be repeated [48]. A large field of interest is the study of optical properties in different semiconductors after ion irradiation with the aim to combine microelectronics and optics for a new generation of monolithically integrated optoelectronic devices. Doping of rare earth elements into semiconductors leads to a remarkable luminescence. Basic investigations were done using Er69Ni31 and Pr87Pt13 alloy LMISs [49]. In the frame of our work we investigated the formation of colour centres in different materials. So very fine structures acting as luminescence sources were formed in LiF crystals by local Ga+ FIB irradiation at 240 K to prevent the impurity diffusion during implantation and handling [50]. Finally, for the determination of lateral strain silicon wafers were locally FIB implanted with Ge and Au ions and investigated by X-ray grazing-incidence diffraction with synchrotron radiation [51]. The evaluated strain is related to the depth distribution of defects within the implanted stripes and corresponds to theoretical predictions. 4. Conclusion The investigation and application of alloy liquid metal ion sources for mass-separated focused ion beams is briefly reviewed with a view to the work of the authors group within the last decade. The use of FIBs with different ion species requires an ion optical column equipped with a mass filter, which makes the system more expensive and complicated and also the handling of the alloy source is somewhat more demanding. However, these disadvantages
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851
are fully compensated by a promising expansion of FIB applications in micro- and nano-technology for tasks which need ions other than gallium. Focused ion beam processing is a serial method with a limited throughput but for R&D, prototyping and fabrication of unique devices which cannot be made by other technologies, an increasing interest and importance in the future are expected. Acknowledgements The author wishes to acknowledge the contributions to this work from Ch. Akhmadaliev, W. Pilz and B. Schmidt from FZD. References [1] J. Orloff, M. Utlaut, L. Swanson, High Resolution Focused Ion Beams, Kluwer Academic/Plenum Press, New York, 2003. [2] P.D. Prewett, G.L.R. Mair, Focused Ion Beams from Liquid Metal Ion Sources, Research Studies Press Ltd., Taunton, Somerset, England, 1991. [3] P.D. Prewett, P.J. Heard, J. Phys. D: Appl. Phys. 20 (1987) 1207. [4] T. Tao, W. Wilkonson, J. Melngailis, J. Vac. Sci. Technol. B 9 (1991) 162. [5] R. Boylan, M. Ward, D. Tuggle, in: Proceedings of the International Symposium for Testing and Failure Analysis, ASM International, Los Angeles, 1989, p. 249. [6] A. Yamaguchi, M. Shibata, T. Hashinaga, J. Vac. Sci. Technol. B 11 (1993) 216. [7] J.C. Gonzalez, M.I.N. da Silva, D.P. Griffis, P.E. Russell, J. Vac. Sci. Technol. B 20 (2002) 2700. [8] L. Bischoff, J. Teichert, V. Heera, Appl. Surf. Sci. 184 (2001) 372. [9] A.J. Steckl, P. Chen, X. Cao, H.E. Jackson, M. Kumar, J.T. Boyd, Appl. Phys. Lett. 67 (1995) 181. [10] S. Reyntjens, R. Puers, J. Micromech. Microeng. 11 (2001) 287. [11] L.W. Swanson, A.E. Bell, G.A. Schwind, J. Vac. Sci. Technol. B 6 (1988) 491. [12] L.C. Chao, B.K. Lee, C.J. Chi, J. Cheng, I. Chyr, A. Steckl, Appl. Phys. Lett. 75 (1999) 1833. [13] L. Bischoff, B. Schmidt, K.-H. Heinig, T. Mu¨ller and S. Hellwig, in: Proceedings of the International Workshop on Nanostructures for Electronics and Optics – NEOP, Dresden, Germany, October 6–9, 2002. [14] J. Fassbender, L. Bischoff, R. Matteis, P. Fischer, J. Appl. Phys. 99 (2006) 08G301. [15] J. Melngailis, J. Vac. Sci. Technol. B 5 (1987) 469. [16] G. Taylor, Proc. R. Soc. A 133 (1964) 383. [17] D.R. Kingham, L.W. Swanson, Appl. Phys. A 34 (1984) 123. [18] A. Wagner, T.M. Hall, J. Vac. Sci. Technol. 16 (1979) 1871. [19] L. Bischoff, J. Teichert, S. Hausmann, T. Ganetsos, G.L.R. Mair, Nucl. Instr. and Meth. Phys. Res. B 161–163 (2000) 1128. [20] L. Bischoff, Ultramicroscopy 103 (2005) 59.
1851
[21] G.L.R. Mair, J. Phys. D: Appl. Phys. 25 (1992) 1284. [22] C.A. Crox, Statistical Mechanics of the Liquid Surface, John Wiley and Son, Chichester, 1980, p. 161. [23] P.D. Prewett, D.J. McMillan, D.K. Jefferies, G.L.R. Mair, Proc. SPIE 393 (1983) 120. [24] W. Knauer, Optik 59 (1981) 335. [25] Y.G. Kim, Y.S. Kim, E.H. Choi, S.O. Kang, G. Cho, H.S. Uhm, J. Phys. D: Appl. Phys. 31 (1998) 3463. [26] G.L.R. Mair, J. Phys. D: Appl. Phys. 29 (1996) 2186. [27] L.W. Swanson, Nucl. Instr. and Meth. Phys. Res. 218 (1983) 347. [28] P. Marriott, Appl. Phys. A 44 (1987) 329. [29] T. Ishitani, K. Umemura, Y. Kawanami, J. Appl. Phys. 61 (1987) 748. [30] L. Bischoff, Ch. Akhmadaliev, German Patent Application DE 10 2007 027 097.8; L. Bischoff et al., Ultramicroscopy 107 (2007) 865. [31] D.G. Brandon, Surf. Sci. 3 (1964) 1. [32] L. Bischoff, E. Hesse, G. Hofmann, F.K. Naehring, W. Probst, B. Schmidt, J. Teichert, Microelectron. Eng. 21 (1993) 197. [33] A.E. White, K.T. Short, R.C. Dynes, J.P. Garno, J.M. Gibson, Appl. Phys. Lett. 50 (1987) 95. [34] L. Bischoff, J. Teichert, E. Hesse, D. Panknin, W. Skorupa, J. Vac. Sci. Technol. B 12 (1994) 3523. [35] E. Hesse, L. Bischoff, J. Teichert, J. Phys. D: Appl. Phys. 27 (1994) 427. [36] S. Hausmann, L. Bischoff, J. Teichert, M. Voelskow, W. Mo¨ller, Jpn. J. Appl. Phys. Part 1 38 (1999) 7148. [37] Ch. Akhmadaliev, L. Bischoff, B. Schmidt, Mater. Sci. Eng. C 26 (2006) 818. [38] Ch. Akhmadaliev, B. Schmidt, L. Bischoff, Appl. Phys. Lett. 89 (2006) 223129. [39] L. Bischoff, Ch. Akhmadaliev, B. Schmidt, Microelectron. Eng. 84 (2007) 1459. [40] M. Posselt, L. Bischoff, J. Teichert, Appl. Phys. Lett. 79 (2001) 1444. [41] M. Posselt, L. Bischoff, J. Teichert, A. Ster, J. Appl. Phys. 93 (2003) 1004. [42] M. Posselt, L. Bischoff, D. Grambole, F. Herrmann, Appl. Phys. Lett. 89 (2006) 151918. [43] D.P. Adams, M.J. Vasile, G. Benavides, A.N. Campbell, Precis. Eng. 25 (2001) 107. [44] P. Sigmund, Phys. Rev. 184 (1969) 383. [45] C. Garcia-Rosales, W. Eckstein, J. Roth, J. Nucl. Mater. 218 (1994) 8. [46] J. Schmidt, H. Tritschler and L. Bischoff, in: Proceedings of the International Conference on Micro and Nano Systems, ICMNS 2002, Kumning, China, August 11–14, 2002. [47] G. Gorbunov, A.A. Levin, E. Wieser, L. Bischoff, D. Eckert, A. Mensch, M. Mertig, D.C. Meyer, H. Reuther, P. Paufler, W. Pompe, Cryst. Res. Technol. 40 (2005) 106. [48] W. Knapp, L. Bischoff, J. Teichert, Appl. Surf. Sci. 146 (1999) 134. [49] L.C. Chao, B.K. Lee, C.J. Chi, J. Cheng, I. Chyr, A. Steckl, J. Vac. Sci. Technol. B 17 (1999) 2791. [50] J. Martin, L. Bischoff, R. Wannemacher, Opt. Commun. 188 (2001) 119. [51] J. Grenzer, U. Pietsch, L. Bischoff, Phys. Status Solidi (a) 202 (2005) 1009.
NIM B Beam Interactions with Materials & Atoms
Nuclear Instruments and Methods in Physics Research B 266 (2008) 1846–1851 www.elsevier.com/locate/nimb
Application of mass-separated focused ion beams in nano-technology L. Bischoff * Research Center Dresden-Rossendorf Inc., Institute of Ion Beam Physics and Materials Research, P.O. Box 51 01 19, D-01314 Dresden, Germany Received 15 November 2007 Available online 15 December 2007
Abstract FIB applications like writing ion implantation, ion beam mixing or ion beam synthesis in the lm- or nm range often require ion species other than gallium. Therefore alloy liquid metal ion sources (LMIS) have to be developed and applied in FIB tools. The energy distribution of ions emitted from an alloy LMIS is one of the crucial parameters for the performance of a FIB column. Different source materials like AuGe, AuSi, AuGeSi, CoNd, ErNi, ErFeNiCr, MnGe, GaBi, GaBiLi, SnPb, . . . were investigated with respect to the energy spread of the different ion species as a function of emission current, ion mass and emitter temperature. Different alloy LMIS’s have been developed and used in the FZD – FIB system especially for writing implantation to fabricate sub-lm pattern without any lithographic steps. Co and various other ion species were applied to generate CoSi2 nano-structures, like dots and wires by ion beam synthesis or to manipulate the properties of magnetic films. Additionally, the possibility of varying the flux in the FIB by changing the pixel dwell-time can be used for the investigation of the radiation damage and dynamic annealing in Si, Ge and SiC at elevated implantation temperatures. Furthermore, a broad spectrum of ions was employed to study in a fast manner the sputtering process depending on temperature, angle of incidence and ion mass on a couple of target materials. These studies are important for the 3D-fabrication of various kinds of micro-tools by FIB milling. Ó 2008 Elsevier B.V. All rights reserved. PACS: 68.03; 61.72; 85.40Ry; 85.45Bz; 71.55i Keywords: Alloy liquid metal ion sources; Energy spread; Focused ion beam; Application; Microstructures
1. Introduction With the invention of the liquid metal ion source (LMIS) in the sixties the focused ion beam (FIB) technique started an impressive development from the laboratory level to high performance industrial equipments [1,2]. Applying these unique point-like sources with a brightness in the order of 106 Acm2 sr1 ion beams with a diameter less than 10 nm and current densities more than 10 Acm2 can be achieved. So a very promising field was opened to use the FIB for mask [3] and integrated circuit repair and modification [4], failure analysis [5] or TEM specimen preparation [6], as well as in material science for sputtering
*
Tel.: +49 351 260 2963; fax: +49 351 260 3285. E-mail address: l.bischoff@fzd.de
0168-583X/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2007.12.008
investigations and applications [7,8], for the formation of quantum dots and wires [9] for the fabrication of MicroElectro-Mechanical Systems (MEMS) [10] and for a lot of other applications in micro- and nano-technology. The majority of FIB systems operating worldwide are suited only for gallium ions, due to the favoured properties, like its low melting point (Tm = 29.6 °C) and the simple emitted mass spectrum. About 99% of the ions in the beam are singly charged. Due to the mass-independent behaviour of electro-static ion optics, no mass filter is needed. The two isotopes have not to be separated and the ion column design becomes easier. But there is an increasing demand for ion species other than Ga. Especially, in order to use the advantages of the FIB, in terms of the high lateral resolution, the flexibility in fluence and pattern design for direct doping purposes, one needs different ion sources to produce the wealth of technologically interesting ions. As
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851
an example B+, As+ or P+ for Si-doping [11] or rare earth elements like Er+, Pr+ and Nd+ for opto-electronic applications [12] metals, like Au+ for plasmonic structures [13] or Fe+, Co+ and Ni+ for selective changes of magnetic properties [14] should be mentioned. To operate LMISs with such source materials is in most cases difficult, due to the high melting points of the pure materials. The composition of suitable binary or ternary alloys and a compatible emitter material can overcome this problem. It should be mentioned, that the application of alloy LMISs requires also an ion optical column equipped with a mass separator to select the wanted ionic species as well as to separate the certain isotopes. The development and preparation, as well as the characterisation of alloy LMISs, and examples of the applications of mass-separated focused ion beams will be summarised in this article. 2. Alloy liquid metal ion sources To focus an ion beam into a spot size of a diameter smaller than one lm a source is needed which emits the ions from a very small area into a limited solid angle. LMISs as well as gaseous field ionisation sources approach these conditions [15]. Because of their broad and versatile applications the following discussion is concentrated on the LMIS only. This type of sources usually consists of an emitter needle with a tip radius of about 5–10 lm, which is covered with a pure metal, i.e. Ga, In, . . . or an alloy, i.e. Au82Si18, Co64Nd36, . . . and heated directly or indirectly. A counter electrode (the extractor) is placed in front of the needle where a high voltage (extraction voltage) in the range of 2–10 kV is applied. At a critical field strength of about some 10 V/nm on the tip a liquid cone, the so called Taylor-cone [16] with a half angle of 49.3° is formed and ion emission occurs from a liquid jet on this cone, mainly due to a field evaporation process [17]. In the following we want to deal more in detail with alloy LMIS. To find a suited material for the desired purpose one have to look for a low temperature melting alloy (mostly in the eutectic composition) with a low vapour pressure and chemical inertness with the source base material. The source preparation itself may be divided in three basic steps: the fabrication of the emitter, the electro-chemical etching using a NaOH solution in the case of tungsten or a mechanical treatment for tantalum emitters to form the tip and the wetting and reservoir filling of the emitter by dipping in a crucible containing the molten alloy in a high vacuum ambient. Table 1 gives an overview about the investigated alloy sources, their melting points and the main application fields. It should be mentioned that all materials exhibit a vapour pressure lower than 108 mbar at the operation temperature beside manganese. A description of the preparation procedure can be found more in detail elsewhere [2,12,18].
1847
Table 1 Alloy composition, melting temperature and application of investigated (alloy) LMIS in the FZD – FIB Material
Melting temperature (°C)
Applications
Au73Ge27
365
Au77Ge14Si9
365
Au82Si18 Co36Nd64
365 566
Co27Ge73 Er69Ni31 Er70Fe22Ni5Cr3 Sn74Pb26 In14Ga86
817 765 862 183 14.2
Mn45Ge54Si1
720
Ga38Bi62 Ga35Bi60Li5
222 250
Au: sputtering, nano-cluster; Ge: Doping, SixGe1x alloy Si: contamination free processing, imaging dto Co: CoSi2 – IBS, Nd: optical + magnetic application Co: magnetic application Er: optical application Fe,Ni,Cr: magnetic application Sn: doping, fundamental research Fundamental source emission investigation Mn implantation in compound semiconductors Ga: acceptor; Bi: shallow donor Li: source for analysis and ion beam lithography All classical FIB applications
Ga
29.7
The primary investigation is the recording of the I-Vcharacteristics from which the onset voltage of the source and its stability can be obtained. Additional the slope of this curve is important for the control adjustment of a stable emission. Analysing the I-V-characteristics as a function of temperature yields the optimum operation point reflecting the behaviour of the surface tension of the melt [19,20]. Normally for emitters, wetted with AuGe or CoNd the extraction voltage decreases with temperature, according to theoretical predictions [21], due to a corresponding decrease in the surface tension coefficient of the alloy. In the case of silicon containing emitters (AuSi, AuGeSi) an entirely different behaviour was found. This behaviour is, in turn, associated with the anomalous behaviour of the surface tension coefficient, arising from a residual crystalline structure at the liquid surface [22]. Therefore, the extraction voltage increases with temperature up to about 800 °C and then decreases as expected slowly at higher temperatures. The energy spread DE of the different ion species is a very important parameter which is significant in determining the final resolution of a FIB arising, in turn, from the chromatic aberrations [2,21]. This parameter was analysed as a function of emission current I, ion mass m and emitter temperature T in the current range of 1–30 lA. For singly charged ions the predicted dependence of the energy spread, DE / I2/3 m1/3 T1/2, found for elementary monoatomic LMIS [24,25] could be reasonable confirmed. For sources, emitting a multiplicity of ion species with different charge states, molecule ions and clusters other dependencies may be obtained [26]. For example, a singly charged Co beam from a CoNd alloy LMIS shows a dependence
1848
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851
Energy spread for differnt ion species I emission = 15 µA
+
Au5
100
+
Au4 +
ΔE FWHM (eV)
Au3 +
Au2
++
Au ++ Er Ni ++ ++ Nd + ++ Co ++Sn + Au ++ Ge Sn Fe + + Ni Ga + Co + + Si Fe ++
++
Si
10
AuGeSi + AuGe
x
fit ΔE ~ m single charged x = 0.33 double charged x = 0.21 cluster ions x = 0.86
+
Li
10
+
100
1000
m/q (amu/e) Fig. 1. Energy spread of singly-charged ions, clusters, and doubly-charged ions as a function of their mass-to-charge ratio obtained from alloy LMIS AuGeSi, AuGe, ErFeNiCr, CoNd, GaBiLi and from pure metal LMIS for Sn and Ga at a constant emission current of 15 lA [20]. The accuracy of the measurements is about 1 eV.
of DE / I0.4 which is in good agreement with some experimental data from [23] and predictions of Mair [26] but in contradiction to the calculations of [24,25]. But the first could be confirmed in the case of an AuSi alloy LMIS. The dependence of DE on the ion mass is presented in Fig. 1. For singly charged ions, DE is proportional to m1/3 which is in good agreement with the theory of Knauer [24] and Kim et al. [25] as well as with the experimental results of Swanson [27]. For doubly charged ions a weaker mass-dependence was found but the absolute values are higher than that of singly charged ones. This behaviour was also found by Marriott [28] and Ishitani
1000
Bi
et al. [29] who indicated the general relationship DE(M2+)/2 < DE(M+) < DE(M2+). For clusters a stronger slope was determined but the data in the literature are rather sparse for a safe comparison. Another interesting point is the composition of the species coming from the alloy wetted emitter. This can be determined by analysing the mass spectrum. A typical high resolution spectrum from an GaBiLi alloy LMIS is shown in Fig. 2 [30]. As can be seen, the single charged ions Bi+ and Ga+ are the dominant species in the beam. For an ion of charge state n, Brandons criterion [31] state that the evaporation field E determines which ion species will
Ga33Bi57Li10 -LMAIS
+
+
Ga
7
Li
+
6
Bi
++
+
Bi4
1
GaBi ++
10
GaBi +
current (nA)
+
Bi 3 Bi 2 Bi 3++
100
+
++
Ga
0.1
50
100
7
+
Li2
200
ExB voltage (V) Fig. 2. Mass spectrum of a Ga45Bi45Li10 alloy LMIS [30].
+
Li
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851
dominate in the beam. If E(I2+) < E(I+), then the surface atom is likely to be field evaporated as a doubly charged ion. But in our case E(I2+) > E(I+) so the single charged ions dominate in the spectrum. 3. Applications of mass-separated FIB The experimental applications, presented here, were performed using partly the in house developed mass-separated FIB system IMSA-100, which is described in detail elsewhere [32], and the modern high resolution FIB equipped with an ion-optical column CANION 31 Mplus (Orsay Physics). As a first example the FIB Ion Beam Synthesis of CoSi2 nanostructures will be described. Buried, single-crystalline CoSi2 layers were produced first by White et al. using ion beam synthesis (IBS), i.e. high fluence Co implantation into a heated silicon substrate followed by a two step annealing [33]. The crucial point was the implantation at elevated target temperatures to avoid amorphisation of the silicon substrate in order to submit the crystalline orientation information to the epitaxial growing silicide. A couple of years ago this process was transferred to the writing FIB technology to obtain smaller structure dimensions [34]. Therefore a Co36Nd64 alloy LMIS was developed [35] and low resistivity, single-crystalline CoSi2-microstructures were successfully fabricated down to dimensions as small as 60 nm [36]. The application of the high resolution FIB admitting a Co2+ beam (60 keV) with a spot size <50 nm opened the possibility to fabricate nanowires (NW) with a width down to 20 nm [37]. Fig. 3 shows a SEM image of such a NW after removing the top Si layer by RIE. But with decreasing NW dimension, the NW growth becomes instable and the NW do not follow the FIB trace in every case. It is always adjusted along the h1 1 0i direction according to energetic reasons. To investigate this phenomenon more in detail lines were written with different ion species to create damage traces. Then Co was diffused into the wafer from the rare side [38] and the resulting NWs looked similar to those in the case of Co FIB implantation. That means, the NW growth is mainly driven by a damage related process, which is more precisely
<110> 20 nm
Fig. 3. SEM image of a 20 nm CoSi2 NW on silicon. The Co implantation fluence was 2.5 1016 cm2 [37].
1849
7 Nd- isotopes not separated 142 143 144 145 146 148 150
27.2 % 12.2 % 23.8 % 8.3 % 17.2 % 5.7 % 5.6 %
fluence variation
Σ = 1x1017 cm-2
Fig. 4. SEM image of CoSi2 NW and nanostructures in silicon after irradiation, Co-diffusion and annealing. Used was a selected Nd2+-FIB where all seven isotopes are separated but simultaneous present on the target after. The over all fluence was 1 1017 cm2.
explained in [39]. Fig. 4 demonstrates a FIB irradiation with a weak adjusted ExB filter using a 60 keV neodymium line resulting in seven small parallel lines corresponding to the natural abundance of the isotopes with an integral fluence of 1 1017 cm2. Depending on fluence and the distance to the neighbour line different shapes and lengths of CoSi2 NWs and nanostructures were obtained. Another problem of the FIB implantation in the nanoscale compared to the broad beam technique (BB) is the current density which can achieve 5–7 orders of magnitude higher values than that of BB implanters and exceeds 10 A cm2 corresponding to a flux of about 1019 ions cm2 s1 in the standing beam. If the spot is very fast moving over a sample surface one gets an averaged flux of the same order as a BB of about 1012 ions cm2 s1. Thus, the crucial parameter adjusting the flux is the pixel dwell time (PDT). This was used for extended investigations of the influence of these extreme flux to the dynamic annealing process after a channelling FIB implantation into Si and SiC [40]. For these experiments a Ge-beam was employed in order to avoid a sample doping or contamination. Varying the PDT (flux) and the target temperature and studying the shape of the implanted profiles by SIMS one can estimate the time scale for the relaxation of complex defects [41]. In both materials, at temperatures of about 200–250 °C significant defect reduction was found within the first 10 ls as well as between 10 ls and 100 s after ion impact. Whereas, at these temperatures the complex defects disappear nearly completely in Si, in SiC some of them survive. Similar investigations were carried out on germanium [42]. A well known and broad field of FIB application is the direct patterning by ion milling also called sputtering [43]. The escape of atoms, ions or clusters from the surface arising from a physical knock-on process caused by an incident ion beam depends on the ion mass and energy as well as the properties of the target material, like structure, density or surface binding energy, and also on the process parameters of temperature and angle of incidence. A theoretical
1850
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851 6
Sputtering yield depending on the ion mass
197
Au
target: Si, incidence angle: 0˚, energy: 30 keV
5
144
Nd
sputtering yield
4
74
3
69 59
Co
Ga
Ge
2
calculated experimental data
28
Si
1
0
0
20
40
60
80
100
120
140
160
180
200
220
ion mass (amu) Fig. 5. Calculated (according to Sigmund [44]) and experimental data of the sputtering yields as a function on ion mass for a 30 keV FIB with single charged ions in silicon for normal incidence.
description of this phenomenon can be found elsewhere [44,45]. The FIB is a very suitable tool for sputtering [1,20] and the characteristic parameter is the sputtering yield. Fig. 5 shows the calculated [44] and experimental data of sputter yields as a function on ion mass for a 30 keV FIB and single charged ions. It is obvious that switching from the usual Ga FIB to gold ions the sputtering efficiency increases by a factor of two which is suitable for the fabrication of micro-tools [46]. On the other hand turning to silicon or lighter ions only weak sputter erosion occurs at normal incidence which is adventurous for low damage imaging and ion lithography applications. In the last section of this review some other selected additional activities using a mass-separated focused ion beam from alloy LMIS will be mentioned to show the capability of this method and also to inspire new applications. Magnetic properties of very small dimensions are of increasing interest for high density storage media. It was observed that in Fe–Cr films the metastable paramagnetic phase can be transformed into a more stable b.c.c. phase which is ferromagnetic. Such a ferromagnetic pattern was written by a fine-focused Cr–ion beam in a paramagnetic alloy (face-centred orthorhombic, 37 at.% Cr) using an alloy liquid metal ion source (Er70Fe22Ni5Cr3). The written pattern were detected by magnetic force microscopy [47]. Furthermore a 60 keV Co2+-FIB was used to change the coupling on a lm-scale in a Ni81Fe19/Ru/Co90Fe10 film from antiferromagnetic to ferromagnetic behaviour [14]. Higher melting point alloys than Ga (i.e. Au73Ge27 and Au77Ge14Si9) can be solidified by quenching. Doing this during operation of an alloy LMIS at which the protrusion on the needle can be frozen a solid nm-tip is formed, which
can be used as an electron emitter after changing the polarity. To regenerate the source this process can be repeated [48]. A large field of interest is the study of optical properties in different semiconductors after ion irradiation with the aim to combine microelectronics and optics for a new generation of monolithically integrated optoelectronic devices. Doping of rare earth elements into semiconductors leads to a remarkable luminescence. Basic investigations were done using Er69Ni31 and Pr87Pt13 alloy LMISs [49]. In the frame of our work we investigated the formation of colour centres in different materials. So very fine structures acting as luminescence sources were formed in LiF crystals by local Ga+ FIB irradiation at 240 K to prevent the impurity diffusion during implantation and handling [50]. Finally, for the determination of lateral strain silicon wafers were locally FIB implanted with Ge and Au ions and investigated by X-ray grazing-incidence diffraction with synchrotron radiation [51]. The evaluated strain is related to the depth distribution of defects within the implanted stripes and corresponds to theoretical predictions. 4. Conclusion The investigation and application of alloy liquid metal ion sources for mass-separated focused ion beams is briefly reviewed with a view to the work of the authors group within the last decade. The use of FIBs with different ion species requires an ion optical column equipped with a mass filter, which makes the system more expensive and complicated and also the handling of the alloy source is somewhat more demanding. However, these disadvantages
L. Bischoff / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1846–1851
are fully compensated by a promising expansion of FIB applications in micro- and nano-technology for tasks which need ions other than gallium. Focused ion beam processing is a serial method with a limited throughput but for R&D, prototyping and fabrication of unique devices which cannot be made by other technologies, an increasing interest and importance in the future are expected. Acknowledgements The author wishes to acknowledge the contributions to this work from Ch. Akhmadaliev, W. Pilz and B. Schmidt from FZD. References [1] J. Orloff, M. Utlaut, L. Swanson, High Resolution Focused Ion Beams, Kluwer Academic/Plenum Press, New York, 2003. [2] P.D. Prewett, G.L.R. Mair, Focused Ion Beams from Liquid Metal Ion Sources, Research Studies Press Ltd., Taunton, Somerset, England, 1991. [3] P.D. Prewett, P.J. Heard, J. Phys. D: Appl. Phys. 20 (1987) 1207. [4] T. Tao, W. Wilkonson, J. Melngailis, J. Vac. Sci. Technol. B 9 (1991) 162. [5] R. Boylan, M. Ward, D. Tuggle, in: Proceedings of the International Symposium for Testing and Failure Analysis, ASM International, Los Angeles, 1989, p. 249. [6] A. Yamaguchi, M. Shibata, T. Hashinaga, J. Vac. Sci. Technol. B 11 (1993) 216. [7] J.C. Gonzalez, M.I.N. da Silva, D.P. Griffis, P.E. Russell, J. Vac. Sci. Technol. B 20 (2002) 2700. [8] L. Bischoff, J. Teichert, V. Heera, Appl. Surf. Sci. 184 (2001) 372. [9] A.J. Steckl, P. Chen, X. Cao, H.E. Jackson, M. Kumar, J.T. Boyd, Appl. Phys. Lett. 67 (1995) 181. [10] S. Reyntjens, R. Puers, J. Micromech. Microeng. 11 (2001) 287. [11] L.W. Swanson, A.E. Bell, G.A. Schwind, J. Vac. Sci. Technol. B 6 (1988) 491. [12] L.C. Chao, B.K. Lee, C.J. Chi, J. Cheng, I. Chyr, A. Steckl, Appl. Phys. Lett. 75 (1999) 1833. [13] L. Bischoff, B. Schmidt, K.-H. Heinig, T. Mu¨ller and S. Hellwig, in: Proceedings of the International Workshop on Nanostructures for Electronics and Optics – NEOP, Dresden, Germany, October 6–9, 2002. [14] J. Fassbender, L. Bischoff, R. Matteis, P. Fischer, J. Appl. Phys. 99 (2006) 08G301. [15] J. Melngailis, J. Vac. Sci. Technol. B 5 (1987) 469. [16] G. Taylor, Proc. R. Soc. A 133 (1964) 383. [17] D.R. Kingham, L.W. Swanson, Appl. Phys. A 34 (1984) 123. [18] A. Wagner, T.M. Hall, J. Vac. Sci. Technol. 16 (1979) 1871. [19] L. Bischoff, J. Teichert, S. Hausmann, T. Ganetsos, G.L.R. Mair, Nucl. Instr. and Meth. Phys. Res. B 161–163 (2000) 1128. [20] L. Bischoff, Ultramicroscopy 103 (2005) 59.
1851
[21] G.L.R. Mair, J. Phys. D: Appl. Phys. 25 (1992) 1284. [22] C.A. Crox, Statistical Mechanics of the Liquid Surface, John Wiley and Son, Chichester, 1980, p. 161. [23] P.D. Prewett, D.J. McMillan, D.K. Jefferies, G.L.R. Mair, Proc. SPIE 393 (1983) 120. [24] W. Knauer, Optik 59 (1981) 335. [25] Y.G. Kim, Y.S. Kim, E.H. Choi, S.O. Kang, G. Cho, H.S. Uhm, J. Phys. D: Appl. Phys. 31 (1998) 3463. [26] G.L.R. Mair, J. Phys. D: Appl. Phys. 29 (1996) 2186. [27] L.W. Swanson, Nucl. Instr. and Meth. Phys. Res. 218 (1983) 347. [28] P. Marriott, Appl. Phys. A 44 (1987) 329. [29] T. Ishitani, K. Umemura, Y. Kawanami, J. Appl. Phys. 61 (1987) 748. [30] L. Bischoff, Ch. Akhmadaliev, German Patent Application DE 10 2007 027 097.8; L. Bischoff et al., Ultramicroscopy 107 (2007) 865. [31] D.G. Brandon, Surf. Sci. 3 (1964) 1. [32] L. Bischoff, E. Hesse, G. Hofmann, F.K. Naehring, W. Probst, B. Schmidt, J. Teichert, Microelectron. Eng. 21 (1993) 197. [33] A.E. White, K.T. Short, R.C. Dynes, J.P. Garno, J.M. Gibson, Appl. Phys. Lett. 50 (1987) 95. [34] L. Bischoff, J. Teichert, E. Hesse, D. Panknin, W. Skorupa, J. Vac. Sci. Technol. B 12 (1994) 3523. [35] E. Hesse, L. Bischoff, J. Teichert, J. Phys. D: Appl. Phys. 27 (1994) 427. [36] S. Hausmann, L. Bischoff, J. Teichert, M. Voelskow, W. Mo¨ller, Jpn. J. Appl. Phys. Part 1 38 (1999) 7148. [37] Ch. Akhmadaliev, L. Bischoff, B. Schmidt, Mater. Sci. Eng. C 26 (2006) 818. [38] Ch. Akhmadaliev, B. Schmidt, L. Bischoff, Appl. Phys. Lett. 89 (2006) 223129. [39] L. Bischoff, Ch. Akhmadaliev, B. Schmidt, Microelectron. Eng. 84 (2007) 1459. [40] M. Posselt, L. Bischoff, J. Teichert, Appl. Phys. Lett. 79 (2001) 1444. [41] M. Posselt, L. Bischoff, J. Teichert, A. Ster, J. Appl. Phys. 93 (2003) 1004. [42] M. Posselt, L. Bischoff, D. Grambole, F. Herrmann, Appl. Phys. Lett. 89 (2006) 151918. [43] D.P. Adams, M.J. Vasile, G. Benavides, A.N. Campbell, Precis. Eng. 25 (2001) 107. [44] P. Sigmund, Phys. Rev. 184 (1969) 383. [45] C. Garcia-Rosales, W. Eckstein, J. Roth, J. Nucl. Mater. 218 (1994) 8. [46] J. Schmidt, H. Tritschler and L. Bischoff, in: Proceedings of the International Conference on Micro and Nano Systems, ICMNS 2002, Kumning, China, August 11–14, 2002. [47] G. Gorbunov, A.A. Levin, E. Wieser, L. Bischoff, D. Eckert, A. Mensch, M. Mertig, D.C. Meyer, H. Reuther, P. Paufler, W. Pompe, Cryst. Res. Technol. 40 (2005) 106. [48] W. Knapp, L. Bischoff, J. Teichert, Appl. Surf. Sci. 146 (1999) 134. [49] L.C. Chao, B.K. Lee, C.J. Chi, J. Cheng, I. Chyr, A. Steckl, J. Vac. Sci. Technol. B 17 (1999) 2791. [50] J. Martin, L. Bischoff, R. Wannemacher, Opt. Commun. 188 (2001) 119. [51] J. Grenzer, U. Pietsch, L. Bischoff, Phys. Status Solidi (a) 202 (2005) 1009.