Recent Progress in Understanding of Lattice Defects in Czochralski-Grown Germanium: Catching-up with Silicon

Recent progress in understanding of lattice defects in Czochralski-

grown germanium: catching-up with silicon

J. Vanhellemont1,a, S. Hens1, J. Lauwaert1, O. De Gryse1, P. Vanmeerbeek1, D. Poelman1, P. Śpiewak2, I. Romandic3 , A. Theuwis3 and P. Clauws1

1Dep. of Solid State Sciences, Ghent University, Krijgslaan 281 S1, B-9000 Ghent, Belgium 2Materials Design Division, Faculty of Materials Science and Engineering, Warsaw University of

Technology, Wołoska 141, 02-507 Warsaw, Poland 3Umicore EOM, Watertorenstraat 33, B-2250 Olen, Belgium

[email protected]

Keywords: germanium, silicon, vacancies, voids, simulation, oxygen, precipitation, FTIR

Abstract. Recent progress is presented in the understanding of grown-in defects in Czochralski-

grown germanium crystals with special emphasis on intrinsic point defects, on vacancy clustering

and on interstitial oxygen. Whenever useful the results are compared with those obtained for silicon.

Introduction

Today the main application of polished germanium wafers is as heavily doped substrate for GaAs

epitaxy in order to produce high efficiency solar cells for space applications. Due to its higher

intrinsic carrier mobility compared to silicon, however, the use of moderately doped germanium

based substrates is being reconsidered for application in advanced nano-electronic devices. 200 and

300 mm germanium (001) oriented dislocation free substrates have recently become available and

germanium on insulator (GeOI) substrates are being developed [1]. For a successful application as

substrate material, these novel germanium based substrates will have to meet the stringent

requirements defined in the SIA road map. One of these requirements is the absence of extended

lattice defects of which the formation depends strongly on the presence and properties of point

defects either introduced during crystal pulling or during wafer processing. In contrast to silicon,

the properties of the intrinsic and extrinsic point defects in germanium are not well-known and a

significant research effort with that respect will be needed in the coming years [2].

Intrinsic point defects in germanium

Mainly three approaches have been used to directly study intrinsic point defect properties in

germanium [3]: quenching from high temperatures followed by annealing at lower temperatures,

plastic deformation at high temperatures and irradiation at low temperatures in order to create

Frenkel pairs and individual point defects. In addition, the study of the diffusion and/or precipitation

of fast diffusing dopants such as Cu and Zn can provide indirect information on intrinsic point

defects. Finally, also ab initio calculations can shed more light on the formation and migration

energy and the electrical activity of the intrinsic point defects.

The quenching technique has been very successful in the assessment of vacancy properties in metals

and has therefore initially also been used extensively for germanium. These preliminary results

demonstrated that quenching of n-type Ge above 800oC resulted in p-type material, pointing to the

creation of quenched-in acceptors. Annealing at 500oC recovered the original n-type, indicating the

out-diffusion/annihilation of the quenched-in thermal acceptors. It was soon realized however that

the early experiments suffered from copper contamination, whereby substitutional copper is a triple

Solid State Phenomena Vols. 108-109 (2005) pp 683-690Online available since 2005/Dec/15 at www.scientific.net© (2005) Trans Tech Publications, Switzerlanddoi:10.4028/www.scientific.net/SSP.108-109.683

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of TTP,www.ttp.net. (ID: 157.193.1.245, Ghent University, Belgium-09/07/13,13:11:43)

acceptor. Further more it soon became clear that quenching experiments cannot yield information on

single vacancies as diffusion is so fast that vacancy clustering during quenching cannot be avoided.

Self-diffusion in Ge is the slowest diffusion process and provides direct evidence of the existence of

an equilibrium concentration of intrinsic point defects. Recent experiments of Giese et al [4]

showed that in contrast to silicon, the self-diffusion coefficient GeD in germanium

(=3.09

2 113.6eV

kTe cm s− − ) is dominated by vacancies. They indeed presented strong experimental

evidence that the vacancy assisted Ge tracer diffusion coefficient *

( ) 0.5V V V

Ge

Ge

C DD

C= is very close to

* *

0.5 0.73V V I I

Ge

Ge Ge

C D C DD

C C= + or * *

I I V VC D C D<< , (1)

which is the opposite of the behavior in silicon. IC

∗ and VC

∗ are the thermal equilibrium self-

interstitial and vacancy concentrations, respectively. CGe is the concentration of germanium lattice

atoms.

Thermal equilibrium concentration of vacancies. Little information can be found on the thermal

equilibrium vacancy concentration *

VC . On the basis of annealing experiments on thermal acceptors

quenched into germanium, Mayburg [5] obtained

2.01

* 23 33 10eV

kT

VC e cm

− −= × , (2)

predicting an equilibrium vacancy concentration of 1.3×1015cm−3 at melt temperature. This is in

good agreement with the estimate of 2.9 − 3.9 ×1015cm−3 that Tweet [6] obtained based on Cu

precipitation and diffusion behavior.

According to Giese et al [4] one can write

3.09

* 24 1 11.18 10eV

kT

V VC D e cm s

− − −= × . (3)

Combining Eq. 2 and Eq. 3 leads to

DV= 3.9e

−1.08eV

kT cm2s−1 . (4)

Fig. 2 shows *

V VC D for germanium and for silicon as a function of the normalized temperature.

Simulation of intrinsic point defect properties. As reliable experimental data for intrinsic point

defects in germanium are still lacking to a large extent, molecular dynamics calculations and

comparison with results obtained for silicon can yield useful information. Already in 1982,

Mashovets [7] published an extensive review on this topic of which the main results concerning

formation and migration energies of self-interstitials and vacancies are represented in Table I

together with more recent results.

Bernstein et al. [8] studied a number of properties of germanium using non-orthogonal tight-binding

(TB) calculations. In case of point defects, these calculations include the vacancy, as well as three

interstitial positions, the tetrahedral, hexagonal, and <110> dumbbell. TB indicated that the lowest-

energy interstitial was the tetrahedral configuration, in contrast with DFT [9] simulations which

684 Gettering and Defect Engineering in Semiconductor Technology XI

found that this was the case for the <110> dumbbell as is the case for Si. Also the Stillinger-Weber

[10] potential was used in order to study self-interstitial structure [11] and from these investigations

also the <110> dumbbell configuration is the most stable one, like in silicon [12]. DFT [9,13]

results show that the <110> dumbbell is significantly more stable in Ge than in Si, and that the

formation energy for germanium vacancy is smaller than the formation energy in silicon vacancy.

These two features explain in part why the vacancy is much more important for self-diffusion in

germanium than in silicon [14,15]. An overview of the results obtained is given in Table I. The most

recent results suggest that the formation energy of self-interstitials is lower in germanium than in

silicon while the migration energy is higher.

Fig. 1: Self-diffusion coefficients in silicon and

germanium normalized to the melt temperature

[4,25].

Fig. 2: *

V VC D for germanium and for silicon,

normalized to the melting temperature Tm.

Void formation in Czochralski germanium

As discussed in the previous paragraphs the vacancy is the dominant intrinsic point defect in

germanium. Dislocation free Cz-grown germanium crystals can therefore contain vacancy clusters

and even macroscopic voids especially when pulled in a hydrogen-rich atmosphere. In high

resistivity dislocation free germanium V2H traps are formed during cooling of the crystal that can

lead to a change of the resistivity in the central part of the crystal (“coring”). In some cases the

excess of vacancies that is created during cooling of the crystal can also lead to the formation of

larger agglomerates that are visible as pits on the wafer surface, comparable to the “COP’s”

observed on silicon wafers. Tweet [6] reported the occurrence of surface pits with a size depending

on the thermal history of the crystal. He observed a pit density reduction when pulling slower or by

reducing the temperature gradient thus keeping the crystal longer at high temperatures or by

postheating the crystal in the puller. The observed pit density reduction was accompanied by an

increase of size of the remaining pits. The observations were explained by a vacancy clustering

mechanism. Four decades later, similar ideas were used to reduce the number of grown-in intrinsic

point defect clusters in large diameter Cz silicon crystals [25].

Nowadays one of the occasionally occurring defects in Czochralski grown germanium is observed

as large surface pits on polished wafers. As illustrated in Fig. 3 these surface pits are of

crystallographic nature and reminiscent of the crystal originated particles (COP's) that are observed

on silicon wafer surfaces where they occur with a much smaller size and higher density than for

germanium. Despite their very low density, this type of defects can be a show stopper for the

Solid State Phenomena Vols. 108-109 685

production of GeOI wafers and for their use for IC processing. A thorough understanding of the

process parameters leading to pit formation is therefore essential.

Assuming that the observed pits are formed by vacancy clustering and vacancy diffusion limited

growth, on can make a crude estimate of the expected size (and density) as for silicon [26,27]

writing for the edge size of an octahedral void

( )0.5Vl D tλ= with ( )

0.50 *1

3

,

8 V V

Si Ge

C C

Cλ π

−=

with CSi = 5x10

22 cm

-3 and CGe = 4.42x10

22 cm

-3 . (5)

In first order approximation one can write assuming the same anneal time

( )( )

0.50

01

Si V VGe Ge

Si Ge V VSi

C C Dl

l C C D

≈ >>

for the same mT

T. (6)

Table I: Calculated and experimental (italic)

formation (EF) and migration (EM) energies in

eV of intrinsic point defects in Si and Ge.

F

VE M

VE F

IE M

IE Ref.

Ge 2.06 - - - [16]

2.07 0.95 - - [17]

1.95 0.98 - 0.44 [18]

1.97 - - - [19]

2.53-

2.63

0.31-

0.4

- 0.09 [20]

1.9 1.0 - - [21]

- - - 0.2 [22]

3.1 - - - [23,24]

- - 6.07 - [11]

1.92 - 2.3 - [9,13]

3.6 - 2.5 - [8]

2.01 1.08 - [5], (4)

Si 2.28 - - - [16]

2.32 1.06 - - [17]

2.13 1.09 - 0.51 [18]

2.38 1.2 - - [19]

2.74-

2.84

0.38-

0.5

- 0.18 [20]

2.35 - - - [21]

- - - 0.22 [22]

4.6 - - - [23,24]

- - 3.76 - [11]

2.66 0.46 3.66 0.94 [12]

3.27 - 3.2 - [9,13]

4 0.4 4 0.9 [25]

Fig. 3: Optical micrographs showing typical

surface pits observed on Ge wafer surfaces.

Fig. 4: Simulated void distribution after a 100 mm

Cz germanium pulling process illustrating the

important impact of the cooling rate of the crystal.

686 Gettering and Defect Engineering in Semiconductor Technology XI

Vacancy clustering during Czochralski growth of germanium crystals can also be simulated using

the clustering software developed for silicon by Sinno et al [28]. This is illustrated in Fig. 4 for a

100 mm crystal pulling process, taking into account the full thermal history of the crystal.

The vacancy solubility and diffusivity data of Mayburg [5] and of Giese et al [4] are used,

respectively. Despite the lack of reliable data on intrinsic parameter properties, the predictions of the

simulation are in good agreement with the observed size distribution of COP's on germanium wafers

[29].

The good agreement that was observed between simulation and observation suggests that the large

pits observed on germanium wafer surfaces are due to vacancy clustering. Due to the much faster

diffusion of vacancies in germanium, the size of the voids in Cz germanium crystals and of the

resulting surface pits on polished wafers, is thus about a factor of ten larger than in silicon crystals

while the density is several orders of magnitude smaller.

Interstitial oxygen in Czochralski germanium

While oxygen is the most important neutral impurity in Czochralski grown silicon, this is not the

case for germanium. One of the reasons is the lower affinity of oxygen for germanium which is

illustrated by the effective segregation coefficient which is about 0.11, indicating that oxygen has

the tendency to stay in the liquid phase. High concentrations of oxygen (1016-10

18 cm

-3) can only be

obtained by growing crystals in an oxygen-containing atmosphere. As for silicon, the normal lattice

site for oxygen is interstitial, whereby it is close to the BC position between two host atoms.

Interstitial oxygen (Oi) is electrically inactive, but can be detected by its infrared absorption bands.

Similar as for silicon, the concentration of Oi is determined by the peak amplitude αmax of the 856 cm

-1 absorption band (T=300 K) [30]

[Oi] (in cm-3) = αmax (in cm

-1) x 5 10

16. (7)

The detection limit at low temperature is about 1015 cm

-3. The IR measurement does not necessarily

reveal the total [Oi] as some of the oxygen may also be present under the form of SiO2 precipitates

that can be detected in infrared absorption by a band around 1100 cm-1. Near the melting point of

germanium, the maximum solubility of interstitial oxygen is about 2.2x1018 cm

-3.

The solubility [30] and diffusivity [31] of interstitial oxygen in Ge are given by

DOI= 0.4e

−2.076eV

kT cm2s−1 and C

OI

* = 2 ×1023e−1.18eV

kT cm−3 (8)

In Fig. 5 the solubility and diffusivity of oxygen in germanium and silicon are compared as a

function of the normalized temperature illustrating that the solubility is very similar while the

diffusivity is much larger in germanium.

Oxygen dimers and oxygen-related shallow donors in germanium. Annealing in the range 300-

500oC leads to the formation of oxygen-related shallow donors (OTDs). At 350

oC one can achieve

an almost complete conversion of Oi into OTDs [32]. It is a reversible process, whereby higher

equilibrium concentrations are obtained at lower annealing temperatures. The formation rate of

OTDs in Ge is much higher than in silicon, in agreement with the higher diffusion coefficient of

oxygen. This is illustrated in Fig. 6 and Fig. 7 showing the rapid decrease of the interstitial oxygen

concentration during a 350°C anneal and the rapid increase of the carrier concentration due to TD

formation [33]. OTDs consist of a family of double donors, with ionization energy ranging from 14

to 18.1 meV for OTD0 and from 31to 40.5 meV for OTD

+. It has recently become clear that dimers

play an essential role in the process of OTD formation [35]. Similar as in silicon, it is believed that

Solid State Phenomena Vols. 108-109 687

every next family member is formed by attachment of an additional oxygen atom. Hereby the

ionization energies become shallower, the larger the cluster is. In other words, the early members

correspond to the deeper states and vice versa.

At the same temperature both the solubility

and the diffusivity of oxygen are much higher

in germanium than in silicon. This might be a

concern when germanium layers are in contact

with a CZ silicon substrate or with oxide

layers during thermal processing. Due to the

higher solubility and diffusivity, germanium

layers will getter oxygen that might give rise

to TD formation during low temperature

treatments or even during device operation.

Fig. 5: Solubility and diffusivity of interstitial

oxygen in germanium and silicon, normalized

with respect to the melting temperature Tm.

Precipitation of oxygen. Interstitial oxygen precipitation in silicon has been studied for nearly half

a century now due to its important impact on material and device properties. Despite the vast

amount of data and analyses available a number of open questions still remain. It is for instance not

yet unambiguously determined what oxide phase is formed nor is the behavior of interstitial oxygen

well understood over the whole anneal temperature range. The quantitative prediction of oxide

precipitate nucleation and its dependence on the history of the material remains a difficult task.

Similar as for silicon, in germanium that is doped with oxygen from the melt up to its solubility,

interstitial oxygen will precipitate during subsequent thermal treatments in the range between 450

and 650°C. This process leads to the formation of germanium oxide precipitates (GeOx) that can be

observed by their typical absorption bands in FTIR spectra. This is illustrated in Fig. 8 showing the

typical spectrum observed after a thermal treatment of a germanium sample with an initial oxygen

content of 2.5 1017 cm

-3. Before the precipitation anneal, the interstitial oxygen concentration was

maximised by performing an oxygen dissolution treatment at 900°C for 5 min in order to dissolve

all oxygen clusters including thermal donors that were formed during crystal growth, followed by a

rapid quench to room temperature using nitrogen gas cooling. Similar as in silicon a shoulder

develops on the interstitial oxygen absorption band which is due to the formation of a germanium

oxide phase. A typical feature is also the presence of a band around 1100 cm-1 that can be attributed

to SiOx precipitates originating from dissolved silicon. As SiO2 is stable at the germanium melt

temperature, these silicon oxide clusters are formed already during crystal growth [30].

Using the technique developed by De Gryse et al [36] for SiOx precipitates in silicon, it is possible

to determine the shape and phase of the GeOx precipitates in annealed oxygen doped germanium.

Fig. 9 shows a simulated GeOx precipitate spectrum superimposed on the experimental one.

Acknowledgement

This work was supported by the Institute for the Promotion of Innovation by Science and

Technology in Flanders (IWT-Vlaanderen) and the Fund for Scientific Research-Flanders (FWO-

Vlaanderen).

688 Gettering and Defect Engineering in Semiconductor Technology XI

Fig. 6: Oi concentration in oxygen doped Ge vs.

annealing conditions [33,34].

Fig. 7: Carrier concentration in oxygen doped

germanium vs. annealing conditions [33,34].

Fig. 8: Typical infrared absorption spectrum revealing

SiOx and GeOx precipitates in oxygen doped

germanium after a heat treatment for 66h at 610°C

followed by 66h at 560°C [34].

Fig. 9: Absorption spectrum of GeOx

precipitates in oxygen doped germanium

with superimposed a simulated spectrum

assuming polyhedral precipitates [34].

References

[1] http://substrates.umicore.com

[2] J. Vanhellemont, O. De Gryse, S. Hens, P. Vanmeerbeek, D. Poelman, P. Clauws, E. Simoen,

C. Claeys, I. Romandic, A. Theuwis, G. Raskin, H. Vercammen and P. Mijlemans: Defect and

Diffusion Forum 230-232 (2004), p. 149

[3] A. Seeger and K.P. Chik: phys. stat. sol. 19 (1968), p. 455

[4] A. Giese, N.A. Stolwijk and H. Bracht: Appl. Phys. Lett. 77 (2000), p. 642

[5] S. Mayburg: Phys. Rev. 95 (1954), p. 38

[6] A.G. Tweet: Phys. Rev. 106 (1957), p. 221

Solid State Phenomena Vols. 108-109 689

[7] T.V. Mashovets: Sov. Phys. Semicond. 16 (1982), p. 1

[8] N. Bernstein, M. J. Mehl, D. A. Papaconstantopoulos: Phys. Rev. B 66 (2002), p. 075212

[9] Antônio J.R. da Silva, A. Janotti, A. Fazzio, R.J. Baierle and R. Motta: Phys. Rev. B 62 (2000),

p. 9903

[10] F.H. Stillinger and T.A. Weber: Phys. Rev. B 31 (1982), p. 5262

[11] H.R. Schober: Phys. Rev. B 39 (1989), p.13013

[12] T. Sinno: Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, MA, 1998

[13] A. Fazzio, A. Janotti, Antônio J.R. da Silva and R. Motta: Phys. Rev. B 61 (2000), p. R2401

[14] W. Frank, N.A. Stolwijk: Mater. Sci. Forum 15-18 (1987), p. 369

[15] H.D. Fuchs, W. Walukiewicz, E.E. Haller, W. Dondl, R. Schorer, G. Abstreiter, A.I. Rudnev,

A.V. Tikhomirov, V.I Ozhogin: Phys. Rev. B 51 (1995), p. 16 817

[16] F. Bailly: in Lattice Defects in Semiconductors (ed. by R.R. Hasiguti),University of Tokyo

Press, Tokyo (1968), p. 231

[17] R. A. Swalin: J. Phys. Chem. Solids 18 (1961), p. 290

[18] K.H. Bennemann: Phys. Rev. 137 (1965), p. A1497

[19] A. Scholz and A. Seeger: Phys. Status Solidi 3 (1963), p. 1480

[20] T. Soma, M. Saeki, and A. Morita: J. Phys. Soc. Jpn. 35 (1973), p.146

[21] J.C. Philips and J.A. Van Vechten: Phys. Rev. Lett. 30 (1973), p. 220

[22] R.R. Hasiguti: J. Phys. Soc. Jpn. 21 (1966), p.1927

[23] V.A. Panteleev: Fiz. Tverd. Tela (Leningrad) 19 (1977), p. 2801 [in Russian]

[24] V.I. Okulich: Author’s Abstract of Thesis for Candidate’s Degree, State University, Gorki

(1964) [in Russian]

[25] V.V. Voronkov and R. Falster: Journ Electrochem. Soc. 149 (2002), p. G167

[26] J. Vanhellemont: J. Appl. Phys. 78 (1995), p. 4297

[27] J. Vanhellemont: Appl. Phys. Lett. 69 (1996), p. 4008

[28] T. Sinno, E. Dornberger, W. von Ammon, R.A. Brown and F. Dupret: Material Science and

Engineering 28 (2000), p. 149

[29] S. Hens, J. Vanhellemont, D. Poelman, P. Clauws, I. Romandic, F. Holsteyns and J. Van

Steenbergen: submitted for publication in Applied Physics Letters.

[30] W. Kaiser and C.D. Thurmond: J. Appl. Phys. 32 (1961), p. 115

[31] J.W. Corbett, R.S. Mc Donald and D.G. Watkins: J. Phys. Chem. Solids 25 (1964), p. 873

[32] P. Clauws: Mater. Sci. Eng. B 36 (1996), p. 213

[33] P. Vanmeerbeek: PhD thesis, Ghent University (2004).

[34] M. De Bock: Master thesis, Ghent University (2002).

[35] P. Vanmeerbeek, P. Clauws and W. Mondelaers: Physica B 308-310 (2001), p. 517

[36] O. De Gryse, P. Clauws, J. Van Landuyt, O. Lebedev, C. Claeys, E. Simoen and J.

Vanhellemont: J. Appl. Phys. 91 (2002), p. 2493

690 Gettering and Defect Engineering in Semiconductor Technology XI

Gettering and Defect Engineering in Semiconductor Technology XI 10.4028/www.scientific.net/SSP.108-109 Recent Progress in Understanding of Lattice Defects in Czochralski-Grown Germanium: Catching-up

with Silicon 10.4028/www.scientific.net/SSP.108-109.683

DOI References

[2] J. Vanhellemont, O. De Gryse, S. Hens, P. Vanmeerbeek, D. Poelman, P. Clauws, E. Simoen, . Claeys, I.

Romandic, A. Theuwis, G. Raskin, H. Vercammen and P. Mijlemans: Defect and iffusion Forum 230-232

(2004), p. 149

doi:10.4028/www.scientific.net/DDF.230-232.149 [6] A.G. Tweet: Phys. Rev. 106 (1957), p. 221

doi:10.1103/PhysRev.106.221 [9] Antônio J.R. da Silva, A. Janotti, A. Fazzio, R.J. Baierle and R. Motta: Phys. Rev. B 62 (2000), . 9903

doi:10.1103/PhysRevB.62.9903 [14] W. Frank, N.A. Stolwijk: Mater. Sci. Forum 15-18 (1987), p. 369

doi:10.4028/www.scientific.net/MSF.15-18.369 [18] K.H. Bennemann: Phys. Rev. 137 (1965), p. A1497

doi:10.1103/PhysRev.137.A1497 [28] T. Sinno, E. Dornberger, W. von Ammon, R.A. Brown and F. Dupret: Material Science and ngineering

28 (2000), p. 149

doi:10.1016/S0927-796X(00)00015-2 [31] J.W. Corbett, R.S. Mc Donald and D.G. Watkins: J. Phys. Chem. Solids 25 (1964), p. 873

doi:10.1016/0022-3697(64)90100-3 [32] P. Clauws: Mater. Sci. Eng. B 36 (1996), p. 213

doi:10.1016/0921-5107(95)01255-9 [36] O. De Gryse, P. Clauws, J. Van Landuyt, O. Lebedev, C. Claeys, E. Simoen and J. anhellemont: J. Appl.

Phys. 91 (2002), p. 2493

doi:10.1088/0953-8984/14/48/367 [2] J. Vanhellemont, O. De Gryse, S. Hens, P. Vanmeerbeek, D. Poelman, P. Clauws, E. Simoen, C. Claeys,

I. Romandic, A. Theuwis, G. Raskin, H. Vercammen and P. Mijlemans: Defect and Diffusion Forum 230-232

(2004), p. 149

doi:10.4028/www.scientific.net/DDF.230-232.149 [9] Antnio J.R. da Silva, A. Janotti, A. Fazzio, R.J. Baierle and R. Motta: Phys. Rev. B 62 (2000), p. 9903

doi:10.1103/PhysRevB.62.9903 [13] A. Fazzio, A. Janotti, Antnio J.R. da Silva and R. Motta: Phys. Rev. B 61 (2000), p. R2401

doi:10.1103/PhysRevB.61.R2401 [15] H.D. Fuchs, W. Walukiewicz, E.E. Haller, W. Dondl, R. Schorer, G. Abstreiter, A.I. Rudnev, A.V.

Tikhomirov, V.I Ozhogin: Phys. Rev. B 51 (1995), p. 16 817

doi:10.1103/PhysRevB.51.16817 [28] T. Sinno, E. Dornberger, W. von Ammon, R.A. Brown and F. Dupret: Material Science and Engineering

28 (2000), p. 149

doi:10.1016/S0927-796X(00)00015-2 [36] O. De Gryse, P. Clauws, J. Van Landuyt, O. Lebedev, C. Claeys, E. Simoen and J. Vanhellemont: J.

Appl. Phys. 91 (2002), p. 2493

doi:10.1063/1.1429800