Yao N. Focused Ion Beam Systems: Basics and Applications

Подождите немного. Документ загружается.

preferential milling had not occurred. It was also reported that the FIB

milling resulted in the formation of numerous other precipitates, which were

not identiﬁed. Hutchison et al.[95] compared EDS and BF TEM analysis of

cellular precipitation in a Cu-Ti binary alloy of lamellae prepared by electro-

polishing and FIB milling. Within the detection limits of EDS, it was

reported that the FIB milling did not result in ‘‘smearing’’ of the elements

between the different phases and that no redeposition occurred. However, BF

TEM indicated that the FIB milling caused an increase in the dislocation

density relative to the electro-polished membranes. De Veirman and Weaver

[96] made a comparative study of preparing lamellae of semiconductor

samples coated with photoresist using tripod polishing, BIB, and FIB milling.

They reported that FIB milling produced the best samples in terms of

homogeneity and thinness while the BIB milling preferentially, sputtered

away the photoresist and the tripod polishing caused the photoresist to be

torn away.

These comparative studies indicate some of the types of artefacts that can

occur when using FIB milling and that these need to be considered when

interpreting the chemical and structural data obtained from FIB prepared

lamellae.

8.6.2 Damage to the top surface due to ion beam

assisted metal deposition

During the FIB preparation of a lamella, metal is deposited over the ROI

using ion beam assisted deposition, in order to planarize the surface of the

sample to prevent the ‘‘curtain effect’’ and to reduce the amount of rounding/

milling of the top of the lamella. (In an FIB/SEM system the Pt may also be

deposited using the electron beam and this is discussed later in the section.)

As the ion beam assisted deposition is an energetic process it can result in

damage, Ga implantation, and intermixing of the depositing species into the

top of the sample. Figure 8.11 shows an EELS map of a Pt strap deposited on

Si using a 30 keV ion beam. The damage from the ion beam breakdown of

the precursor has been studied by numerous authors [97–99]. Walker and

Broom [98] investigated the resulting depth of damage for Si, In P, and GaAs

and reported depths of damage of 30, 20, and 20 nm, respectively. Detailed

Auger and PEELS/EDS analysis of Pt layers deposited on TiN and W have

shown that the interface consists of two bands, a brighter band that is carbon

rich and a darker band that is an intermixed zone with the substrate [63].

Kempshall et al.[99] studied the effects of reducing the incident energy to

5 keV and using e-beam induced deposition of Pt in an FIB/SEM system.

Focused ion beam systems238

They reported that reducing the ion beam energy from 30 to 5 keV, decreased

the depth of damage in Si from 50 nm to 13 nm and that no damage beneath

the e-beam deposited Pt layer could be observed. The e-beam deposition rate

is dependent on the incident energy of the electrons, i.e., the number of SE

generated, to break down the precursor gas and thus can be increased by a

factor of ten by reducing the electron beam energy from 30 keV to 5 keV.

However, if an electron beam of 400 pA at 5 keV is used the deposition rate is

still ten times slower than that when using a 300 pA, 30 keV ion beam.

Therefore, to deposit the Pt strap the ﬁrst 100 nm can be deposited using

e-beam deposition and the rest completed using ion beam deposition [89].

The simplest method to prevent ion beam damage to the surface of a

sample if using a single beam system is to coat the surface with a 100 nm thick

protective ﬁlm prior to placing it into the FIB system. A range of materials

have been used for this purpose, including Au [100], plasma-polymerized

layers [101], and amorphous silicon [102]. A disadvantage of this is that as the

whole sample is coated it may prevent its use for subsequent processing and

device fabrication.

8.6.3 Ga ion implantation and sidewall damage

During the milling of a TEM lamella, steady state sputtering conditions will

be reached in which the rate of Ga implantation balances the rate of Ga

sputtering. Ishita et al.[102] modeled ion retention during steady state

sputtering and reported a good correlation between their model and

experimental results. They used EDS to determine the Ga concentration in a

lamella prepared using a 30 keV ion beam. On the assumption that the

30 nm

Silicon

α Silicon

Pt strap

Figure 8.11 EELS image of a Pt strap on silicon.

Preparation for physico-chemical analysis 239

implanted Ga was within the ﬁrst 10 nm of the sidewalls, the authors calcu-

lated the Ga concentration in silicon and tungsten layers as 4 and 9 at.%,

respectively. Langford and Petford-Long et al. [103] investigated the effect

of different ion beam milling conditions on the implanted steady state Ga

concentration in silicon lamellae. They reported that the Ga concentration

decreased by a factor of 2.5 when the energy of the FIB was reduced from

30 to 5 keV. Lamellae milled using iodine gas assisted etching (GAE) with

30 and 5 keV ion beams were reported to contain 2.7 and 3 times less Ga

than those prepared using a 30 keV ion beam. The Ga concentration has

also been investigated using APFIM. Martens et al.[104] analyzed the

distribution in FIB prepared AP tips and found that over the ﬁrst 20 nm the

Ga concentration was 30 at.%. Ferryman et al.[105] investigated the Ga

distribution in silicon lamellae using XPS and reported that the Ga con-

centration varied over the lamella, and attributed this to the possible taper

of the lamella. Ga ions implanted into the sidewalls of a lamella have been

reported to coalesce in some materials to form Ga rich regions. Tanaka

et al.[106] used a TEM that had been modiﬁed to incorporate an FIB to

show that in GaAs the implanted Ga precipitated as liquid or amorphous

particles of 200 nm diameter because of the selective etching of As, while

in Si, SiO

, and SiC samples it precipitated as small dots with a diameter of

1–10 nm.

The depth of the damage at the sidewalls of FIB prepared TEM lamellae

has been investigated for a wide range of materials [107–110]. To do this cross

sections of the lamellae and cross sections of the ends of milled trenches have

been made. Kato et al. [109] and Sutton et al.[110] investigated the effect of

different beam currents on the depth of the damage in Si, GaAs, and InP.

Both authors reported that over the range of 100 to 500 pA the depth of

damage was independent of the magnitude of the beam current.

In addition to using TEM, scanning probe microscopy (SPM) has also

been used to assess FIB induced artefacts. Huey et al.[111] and Basnor et al.

[112] used AFM (atomic force microscopy) to characterize the swelling in

silicon when implanting with low doses of Ga. They both reported that for

doses above 10

and below 10

Ga/cm

, swelling of the implanted region

occurred. Brezna et al.[113] used scanning capacitance microscopy (SCM) to

measure the FIB damage beneath FIB milled trenches in silicon. They

reported that the electrical properties of the silicon were affected by distances

up to 620 nm beneath the FIB milled regions.

Typically FIB prepared TEM lamell ae ar e of the order of 100 nm thick

and this when coupled with the damage at the sidewalls may affect HREM.

Therefore, many methods have been used to reduce both the thickness of

Focused ion beam systems240

thelamellaeandthedamageatthesidewalls.Themostcommonpracticeis

to reduce the energy of the FIB [11 4] (t o typically b etwe en 1 and 5 k eV) and

to raster this over the sidewal ls of the lame llae as they a re t ilted t o between

±5 and 10



. ( Th e lin e m il ling pr oc ed ure ca n n ot b e u sed du e t o th e i n cr eas e

in thespot sizeand the decreaseinthesignal t o nois e ra tio.) Unfortuna tel y,

as the lamella is no t rotated an d becau se of the large incident angles pre-

ferential sputtering may oc cur during this low energy (LE) milling [11 5, 116].

The amo un t wi l l d ep en d on t he sp ut ter y iel ds of t he different materials, the

angl e o f til t us ed, and h ow much ma ter ial is s putte red away fr om the

sidewalls. I n Figure 8.12(a) a BF TEM image of a c ross secti on o f a CdS

solar cell lamel la which w as LE FIB m il led u sing a 5 0 p A, 5 keV beam at 10



for 2 minutes on each s ide s hows that preferenti al milli ng of the di fferent

la yer s h as oc cur re d. ( Th es e co nd iti ons were determined to sputter

ap pr ox imate l y 70 n m fro m e ith er sid e of the lamella.) For c omparison,

Figure 8.1 2(b)shows theequivalentBFTEMcrosssectionofalamella

pr ep are d u sin g a 10 0 pA bea m a t 30 kV an d a t i lt of 1 . 5



du rin g th e lin e

milling thinning.

The suitability of FIB prepared lamellae for HREM has also been

improved by making angle cuts across the top edge of the lamella [117]

(Figure 8.13(a)). For silicon lamellae the apex of the resulting taper will be

completely amorphous but as the distance from the apex is increased, an

optimum region in terms of the thickness of the crystalline region and

damaged layer will occur for HREM. The angle cuts may also be made using

an LE FIB beam to further improve the suitability of the lamellae for

HREM. Figure 8.13(b) shows a BF TEM image of an angled cut lamella.

Tapered regions in the vertical direction may also be prepared by tilting

the lamellae to large angles [69,118](4



and above) during the line milling

(b)

CdS

ITO

100 nm

(a)

Figure 8.12 BF TEM images of a CdS solar cell which has been prepared

using (a) LE FIB milling and (b) standard FIB milling.

Preparation for physico-chemical analysis 241

procedure. Furthermore, milling at large angles is also reported to reduce the

depth of the damage at the sidewalls [119].

GAE involves the exposure of an area to a reactive gas during FIB milling.

It is reported to increase the etching rate, reduce the redeposition of material,

and decrease the number of crystal defects and amount of Ga incorporated

into a sample [120]. Sugimoto et al. [121] used chlorine GAE to reduce the

damage in their GaAs samples and Yamaguchi and Nishikawa [122] used

iodine GAE to improve InP cross section specimens for HREM. Yamaguchi

and Nishikawa [122] reported that GAE reduced the thickness of the damage

layer from 31 nm to 2–5 nm and prevented the formation of some FIB

artefacts such as the occurrence of micro-crystals. Langford and Petford-

Long [103] and Kato et al. [109] investigated the use of iodine GAE on silicon

cross sections and although it increased the sputter yield, both reported that

it did not reduce the thickness of the amorphous layer.

BIB milling, which enables lower incident energies to be used than is

currently possible in FIB systems (if a ﬂoating gun is used the energies can be

as low as a few hundred electronvolts), has been used to reduce the FIB

damage from Si p-n junctions [123] and steel specimens [124]. Kato et al.[109]

reported that BIB milling of FIB prepared silicon lamellae, using a 4 keV

beam at 4



, reduced the depth of the damage at the sidewalls of FIB milled

pillars from 30 nm to 8 nm. When post-BIB milling FIB prepared lamellae the

mounting of the sample in the miller is important to prevent redeposition

(a)

Angled cut in silicon

(b)

2 µm

Figure 8.13 (a) Schematic of FIB angle milling to produce a tapered

specimen and (b) BF TEM of a lamella after an angle cut.

Focused ion beam systems242

onto the lamellae. Figure 8.14(a) illustrates the ideal mounting of an in-situ

lift-out or an ex-situ lift-out sample for BIB milling on a 200–300 nm thick

support membrane such that the top and bottom sides of the lamella can be

milled simultaneously while it is being rotated. All the edges and surfaces

close to the lamella should be at a shallow angle of incidence relative to

the incident beam to prevent sputtering and redeposition onto the lamella.

Figure 8.14(b) shows HREM lattice images of InP lamellae prepared using

(i) the standard milling conditions, (ii) iodine GAE, and (iii) LE (5 keV) FIB

milling, viewed along the <110> zone axis. Comparing the lattice images

indicates the beneﬁts of using GAE and post-BIB milling. The relative

improvement of the lattice image of the GAE lamella will partly be due to the

decrease in the thickness of the damage at its sidewalls, while the relative

improvement of the lattice image of the LE FIB milled lamella will be due to

both a decrease in its thickness and the depth of the damage at the sidewalls.

Cross s ections of these three lamellae showed that the thickness of the amor-

phous l ayers were 15 nm for the lamellae pr epared using the standard FIB

milling conditions and 3–4 nm for t he lamellae prepared using the GAE and

LE FIB m illing.

Another method [125] that has been used so that ex-situ lift-out prepared

lamellae can be BIB milled has consisted of depositing Pt straps from the

corners of the lamella to the edges of the support grid such that these straps

(ii)

(iii)(i)

BIB direction

(a)

(b)

Figure 8.14 (a) Mounting of a sample for LE BIB milling to remove the FIB

damage and (b) lattice images of InP lamellae prepared using (i) standard

milling, (ii) GAE, and (iii) LE FIB milling.

Preparation for physico-chemical analysis 243

support the lamella once the carbon support membrane has been milled away

during the BIB milling.

8.6.4 Redeposition and preferential milling

In addition to the damage generated at the sidewalls, two other FIB artefacts

that may occur are redeposition onto the lamellae and preferential milling

[115,126–128]. The amount of redeposition will be dependent on the geo-

metry of the FIB cuts and the milling regime used to thin the lamella. Cairney

and Muroe et al.[128] reported that sequential line milling reduces, but does

not completely eliminate, redeposition and that LE milling does not affect the

amount of redeposition at the sidewalls of lamellae prepared without using

the line milling procedure.

Preferential milling of samples composed of materials with different

sputter rates will inevitably occur and the amount will be dependent on the

differences in the sputter yields and the structure and geometry of the

samples. Figure 8.12(b) shows a BF TEM image of a cross section through

an FIB lamella of a CdS solar cell prepared using standard FIB milling

conditions, which shows that preferential milling of the different layers has

occurred. The line milling step was performed using a beam current of

100 pA and at a tilt angle of 1.5



. Figure 8.15 shows an AFM image of an

FIB cross section through a nanoindentation into a 4 mm thick PVAC spun

onto silicon. The cross section was milled using a 30 kV, 100 pA beam cur-

rent without any tilting during the line milling procedure. The image shows

that relative to the silicon the PVAC has been recessed in places by up to

200 nm. In addition to affecting quantitative chemical analysis, preferential

02040

400

200

µm

Figure 8.15 AFM image of an FIB cross section through the center of a

nano-indent into a 4 mm thick PVAC polymer ﬁlm on silicon.

Focused ion beam systems244

milling will also place a limitation on the minimum thickness that a specimen

can be prepared.

8.7 Summary

In this chapter the different FIB techniques used for preparing specimens for

physical and chemical analysis have been discussed. The diverse range of

examples given illustrates the versatility of the different techniques and the

considerable impact that they have had and are having on material/device

analysis. It can only be expected that the range of novel uses and applications

will continue to rapidly increase.

References

[1] R. M. Langford, S. O’Reilly and I. Mewn. Mater. Res. Soc. Symp. Proc., 739

(2002), 65.

[2] Y. Katoh, M. Ando and A. Kohyama. Nucl. Mater., 323 (2003), 251–62.

[3] D. R. G. Mitchell, D. J. Attard, G. A. Collins and K. T. Short. Coatings Tech.,

165 (2003), 107–18.

[4] H. Saka, K. Kuroda, M. H. Hong et al. ICEM, 13 (1994), 1009.

[5] D. M. Longo, J. M. Howe and W. C. Johnson. UltraMicroscopy, 80 (1999), 69.

[6] A. J. Smith, P. Munroe and T. Tran. J. Mater. Sci., 36 (2001), 3519.

[7] C. Burmer, S. Gorlich and S. Pauthner. Microelectronics Reliability, 38 (1998), 987.

[8] G. Mardingly. Inst. Phys. Conf. Ser., 164 (1999), 575.

[9] S. Tsuji, K. Tsujimoto and H. Iwama. IBM J. Res. Develop., 42 (1998), 509.

[10] J. C. Rei ner, P. Gasser and U. Sennhauser. Microelectronics, 42 (2002), 1753.

[11] S. Mo rris, S. Tatti , E. Black, N. Dickson et al. Proc. 17th Int. Symp. Testing and

Failure Analysis, (1991), 471.

[12] D. Verlleij. Microelectronics Reliability, 38 (1998), 869.

[13] D. J. Larson, A. K. Petford-Long, A. Cerezo and G. W. D. Smith. Acta Mater.,

47 (1999), 4019.

[14] J. M. Cairney, P. R. Muroe and J. H. Schneibel. Scripta Mater., 42 (2000), 472.

[15] R. Pantel, G. Auvert and G. Mascarin. Microelectronic Eng., 37/38 (1997), 49.

[16] B. Wei, K. -R. Philipp, U. Bader et al. UltraMicrosc., 85 (2000), 93.

[17] E. Zschech, E. Lang er, H. -J. Engelmann and K. Dittmar. Mater. Sci. Semicond.

Process., 5 (2003), 457.

[18] H. White, Y. Pu, M. Rafailovich, R. Sokolov, A. H. King et al. Polymer, 42

(2001), 1613.

[19] J. Loos, K. J. Jeroen, F. Morrissey and R. A. J. Janssen. Polymer, 43 (2000), 7493.

[20] K. Hosi, S. Ejiri, W . Probst et al. J. Microsc. , 201 (2001), 44.

[21] H. Engqvist, J. -E. Schultz-walz, J. Loof et al. Biomaterials, (2003).

[22] R. Haswell, D. W. McComb and W. Smith. J. Microsc., 211 (2003), 161.

[23] P. J. Heaney, E. P. Vicenzi, L. A. Giannuzzi and K. J. T. Livi. Am. Mineral., 85

(2001) 1094.

[24] B. I. Prenitzer, L. A. Giannuzz i, K. Newman et al. Mater. Trans. A, 29A (1998),

2399.

Preparation for physico-chemical analysis 245

[25] F. Shaapur, T. Stark, T. Woodward and R. J. Graham. Mater. Res. Soc. Proc.,

480 (1997), 173.

[26] D. P. Basile, R. Boylan, B. Baker, K. Hayes and D. Soza, Mater. Res. Soc. 254

(1992), 23.

[27] E. Kirk, D. Williams and H. Ahmed. Inst. Phys. Conf. Ser., 100 (1989), 501.

[28] K. Park. Mater. Res. Soc. Proc., 199 (1990), 271.

[29] T. Ohnishi, H. Koike, T. Ishitani et al. Proc. 25th Int. Symp. Testing and Failure

Analysis, 449 (1999) .

[30] B. W. Kempsall, B. W. Schwarzx and L. A. Giannuzzi. ICEM, 249 (Durban,

South Africa, 2002).

[31] F. Altmann and D. Katz er. Thin Solid Films, 344 (1999), 609.

[32] D. M. Schraub and R. S. Rai. Prog. Cryst. Growth Charact. Mater., 36 (1999), 99.

[33] R. M. Anderson and S. D. Walck. Mater. Res. Soc. Proc., 480 (1997), 187.

[34] FEI service notes.

[35] S. J. Klepies, J. P. Benedict and R. M. Anderson. Mater. Res. Soc., 179 (1988).

[36] L. R. Harriott. J. Vac. Sci. Technol. B, 11 (1993), 2200.

[37] R. J. Young, J. R. A. Cleaver and H. Ahme d. J. Vac. Sci. Technol. B, 11 (1993), 234.

[38] J. Szot, R. Hornsey, T. Ohnishi and S. Minagawa. J. Vac. Sci. Tech. B, 10

(1992), 575.

[39] T. Ishitani and T. Yaguchi. Microsc. Res. Tech., 35 (1996), 320.

[40] A. Gandhi and J. Orloff. J. Vac. Sci. Technol. B, 8 (1990), 1814.

[41] D. Santamore, K. Edinger, J. Orloff and J. Melngailis. J. Vac. Sci. Tech. B , 15

(1997), 2346.

[42] W. Boxleitner, G. Hobler, V. Kluppel and H. Cerva. Nucl. Instrum. Meth. Phys.

Res. B, 175 (2001), 102.

[43] J. F. Walker. Inst. Phys. Ser., 157 (1997), 469.

[44] G. Deshais and S. B. Newcomb. Eurem 12, 573 (Brno, Czec h Republ ic, 2000).

[45] Y. Z. Huang, S. P. Lozano-Perez, R. M. Langford, J. M. Titchmarsh and M. L.

Jemkins. J. Microsc., 207 (2002), 129.

[46] M. H. Overwijk, F. C. van den Heuvel and C. W. T. Bulle-Lieuwma. J. Vac. Sci.

Technol. B, 11 (1993), 2021.

[47] J. Ebel, C. Bozada, T. E. Schlesinger, C. Cerny et al. IEEE 36th Annual

International Reliability Physics Symposium, (Reno, Nevada, 1996).

[48] A. I. Oliva, A. Romero, J. L. Pena, E. Anguiano and M. Aguilar. Rev. Sci.

Instrum., 67 (1996) , 1917.

[49] K. S. K. Kim, S. C. Lim, I. B. Lee et al. Rev. Sci. Instrum., 74 (2003), 9.

[50] F. A. Stevie, T. C. Shane, P. M. Kahora et al. Proc. Symp. Applied Surface

Analysis, 23 (Bu rlington, MA, USA, 1995) 61.

[51] L. R. Herlinger, S. Chevacharoenkul and D. C. Erin. Proc. ISTFA’96, 199 (1996).

[52] R. M. Langford and Petford-Long. J. Vac. Sci. Tech. A, 19 (2001), 982.

[53] R. M. Langford and C. Clincton. Micron, 35:7 (2004), 607.

[54] A. Hall, W. G. Matthews, R. Superﬁne, M. R. Flavo an d S. Wasburn Appl.

Phys. Lett., 82(15) (2003), 2506.

[55] P. Avouris, T. Hertel, R. Martel et al. Appl. Surf. Sci . 141 (1999), 201.

[56] A. E. M. De Veirman. Mater. Sci. Eng., B00 (2003), 1.

[57] D. M. Longo, J. M. Howe and W. C. Johnson. UltraMicroscopy, 80 (1999), 69.

[58] Y. Yabuuchi, T. Okano, S. Tametou, M. Arai and T. Kouzaki. Eurem 12, 565

(Brno, Czech Republic, 2000).

[59] M. Saito, T. Aoyama, T. Hashimoto and S. Isakozawa. Jpn. J. Appl. Phys., 37

(1998), 355.

Focused ion beam systems246

[60] T. Ishitani, Y. Taniguchi, S. Isakozawa et al. J. Vac. Sci. Technol. B, 16 (1998),

1907.

[61] T. Yaguchi, T. Kammino, H. Kobayashi, H. Koike, K. Tohji, K. Nakatsuka

and R. Urao. Eurem 12, 569 (Brno Czech Republic, 2000).

[62] R. J. Young, E. C. G. Kirk, D. A. Williams and H. Ahmed. Mater. Res. Soc.

Proc., 199 (1990), 205.

[63] R. M. Langford, Y. Huang, S. Lozano-Perez, J. M. Titchmarsch and A. K.

Petford-Long. J. Vac. Sci. Technol., B19 (2001) 755.

[64] H. Bender. Inst. Phys. Conf., 164 (1999), 595.

[65] T. Kato, T. Murago, Y. Iijima et al. J. Electron. Microsc., 53 (2004), 501.

[66] T. Tsutsumi, E. Suzuki, K. Ishii et al. J. Vac. Sci. Technol. B, 17 (1999), 1897.

[67] S. B. Newcomb and W. A. J. Quinton. Eurem 12, 575 (Brno, Czech Republic

2000).

[68] R. M. Langford, D. Ozkaya, D. Zhou, A. K. Petford-Long and C. Stanley.

Proc. ICEM, 251 (Durban, 2002).

[69] K. Minow a, K. Takeda, S. Tomimatsu and K. Umemura. J. Crys. Growth, 210,

(2000), 16.

[70] C. Dix, P. F. McKee, A. R. Thurlow et al. J. Vac. Sci. Technol. B, 12 (1994), 3708.

[71] J. M. Cainey, P. R. Munroe and M. Schneibel. Scripta Mater., 42 (2000), 473.

[72] A. Heilmann, F. Altmann, D. Katzer et al. Appl. Surf. Sci., 144 (1999) , 628.

[73] S. V. Prasad, J. R. Michael and T. R. Christenson. Scripta Mater., 48 (2003), 255.

[74] R. E. Dunin-Borkowski, S. B. Newcomb, M. R. McCartney, C. A. Ross and

M. Farhoud. Electron Microscopy and Analysis, Institute of Physics Conference

Series, 168 (2001), 485.

[75] S. M. Scharz, B. W. Kemps hall, L. A. Giannuzzi and M. R. McCartney, Microsc

and Microanal., 9 (suppl. 2) (2003), 116.

[76] H. Ximen and P. E. Russell. UltraMicroscopy, 42 (1996), 1526.

[77] A. Olbrich, B. Ebersberger, C. Bolt et al. J. Vac. Sci. Technol. B, 17 (1999), 1570.

[78] C. Menozzi, C. G. Gazzadi, A. Alessandrini and P. Facci, UltraMicroscopy ,

104:3–4 (2005), 220.

[79] K. Akiyama, T. A. Eguchi, Y. Fujikawa et al. Rev. Sci. Instr., 76:3 (2005), 33705.

[80] D. J. Larson, A. K. Petford-Long, A. Cerezo et al. Appl. Phys. Lett., 73 (1998), 8.

[81] D. J. Larson, D. T. Foord, A. K. Petford-Long, A. Cerezo and G. W. S. Smith.

Nanotechnology, 10 (1999), 45.

[82] D. J. Larson, B. D. Wissman, R. L. Martens et al. Microsc. Microanal. 7 (2001) 24.

[83] M. K. Miller, K. F. Russell, G. B. Thompson. Ultra Microscopy, 102(4) (2005),

287.

[84] H. Z. Wu, S. G. Roberts, G. Mobus and B. J. Inkson. Acta Mater., 51 (2003), 149.

[85] R. M. Langford, G. Dale, P. J. Hopkins, P. J. S. Ewen and A. K. Petford-Long.

J. Micromech. Microeng., 12 (2002).

[86] B. J. Inkson, T. Steer, G. Mobus and T. Wagner. J. Microsc., (2002).

[87] T. Sakamoto, Z. Cheng, M. Takahashi, M. Owari and Y. Nihei. Jpn. J. Appl.

Phys., 37 (1998), 2051.

[88] H. Ximen, R. K. Defreez, J. Orloff, J. Elliott et al. J. Vac. Sci. Technol. B, 8

(1990), 1361.

[89] R. Kruger. Micron, 221 (1999), 30.

[90] P. Gnauck and P. Hoffrogge. Proc. SPIE: Reliability, Testing and

Characterization of MEMS/MOEMS II, 4980 (2003), 106.

[91] D. J. Barber. Ultr aMicroscopy, 52 (1993), 101.

Preparation for physico-chemical analysis 247