Implantation-induced structural defects in highly activated USJs: Boron precipitation and trapping in pre-amorphised silicon

F C 't' 1'* ZE 1,2yQ. l,3yS · 14 4 · 4 2 . n� l�no , .' ssa 5' . 1U '6 ' plege ,F.Torregrosa, 1. Duchame , P.Bou1enc ,

C.Tavermer , O.CoJocaru , D.B1avette , D.Mange1inck7, P.F.Fazzinis, M.Quillec9, M.BazizilO, M. Hack enberg II , S. Boninelli 12

1 CNRS-LAAS, Univ. of Toulouse, 7 avo du Col. Roche, 31400 Toulouse France 2 STMicroelectronics 850 rue 1.Monnet F-38926 Crolles France

3 CEMES-CNRS, 29 rue 1. Marvig 31055 Toulouse France 4 lBS avo G.lmbert 13790 Peynier Rousset France

6 •

5 Max-Planck-Institut fLir Eisenforschung, Max-Planck-Str. 1, 40237 Dusseldorf, Germany Umv de fou�n, G��, U� CNRS 6634 BP 12, Av de l'Universite 76801 Saint Etienne de Rouvray, France

Umverslte Paul Cezanne, TM2NP- UMR 6242 CNRS Case 142, 13397 Marseille France LPCNO-TNSA, 135, avenue de Rangueil, F-31077 Toulouse France

9 '

Probion, 37 rue de Fontenay, 92220 Bagneux, France 10 Globa�foundries, Wilschdorfer Landstrasse 101, 01109 Dresden, Gennany

FhG-TTSB, Schottkystrasse 10, 91058 Erlangen Germany 12 .

' MATTS-TMM-CNR, VIa Santa Sofia 64, T-95123 Catania, Italy

*Tel: +33 5 61 33 62 54, Fax: +33 5 61 33 62 08, email: fuccio@laas.fr

Abstract

Today, most of the state-of-the-art USJs fabrication processes involve the formation of an amorphous surface layer before or during the dopant implant step. In this paper, we present a review of some recent experimental studies on the Boron precipitation and trapping in pre-amorphised USJs. These studies suggest that the physical mechanism governing the Boron trapping mainly depends on the Boron concentration left below the a/c interface after the implant. Tn addition to providing a contribution to the understanding of the Boron trapping phenomenon, these results clearly indicate that physical models for the fonnation of large Boron precipitates need to be implemented in TCAD simulators for a comprehensive description of the USJ fabrication process.

1. Introduction

To continue scaling down complementary metal-oxide semiconductor devices, ultra-shallow source/drain junctions (USJs) with low resistivity must be fabricated. Today, most of the state-of-the-art USJs fabrication processes involve the formation of an amorphous surface layer before or during the dopant implant step, with the inherent advantage of reducing ion channeling during implantation and improving dopant activation. Tn the case of Boron p + -n junctions, these methods range from Ge+[I] or low-temperature

97S-1- 4673-1257-S/12/$31.00 ©20121EEE

C+ [2] pre-amorphisation, through F+ co-implantation [3], molecular BF/ and cluster implants [4], to BF3 plasma implantation technology [5].

Amorphising implants create a large amount of damage in the Silicon lattice just beyond the amorphous/crystalline (a/c) interface, which upon annealing precipitates into the so-called end-of-range (EOR) defects. Tn such conditions, it is known that, after annealing, the tail of the Boron concentration profiles shows an anomalous peak [6] in correspondence with the EOR defects, which is generally associated the "trapping" of Boron by the defects. Tn addition, it is generally accepted that the boron atoms contained in the so-called "trapping peak" are immobile and electrically inactive. However, the intensity and annealing behavior of the observed trapping peak can drastically vary depending on the process conditions, especially when considering that extremely high supersaturations of dopant atoms (Boron), substrate self-interstitial atoms (Siint) and co-implanted impurities (Carbon, Fluorine) can be simultaneously introduced during the implant. Precipitation of these supersaturated species occurs during anneal and results in the formation of Boron clusters and/or extended Si-interstitial defects, which can both be responsible for Boron trapping through direct Boron precipitation or decoration of extended Si-interstitial defects [7], respectively.

Tn this paper we will therefore present a review of some recent experimental studies that contribute to improving the understanding of this phenomenon. Tn

" • - • after 1000s 1 E2l ,

• - • after 100, ,

M !

I • - • after 30s , ,

lE20 • • s � lE19 ;: 8 "

� lE18 " e o (I) 1 E17 ,

Oepth (micron)

Figure 1 - Boron trapping at EOR defects in Ge+

pre-amorphised USJs (SIMS concentration profiles) during

annealing at 800De. B+ implant: 0.5 keY, Ixl015 cm_2.

particular, we will compare Boron trapping experiments obtained in the presence of both high supersaturations of Boron and Fluorine (typical of BF iBF 3 implants) with reference experiments obtained in pre-amorphised Boron USJs. Finally, the results presented in this paper are also expected to provide a background for the further improvement of the physical models used in TCAD process simulations.

2. Boron trapping in Ge+ pre-amorphised USJs

We first consider the case in which the Boron concentration at the alc interface is low after the implant «lxlO18 cm-3). For instance, after a 0.5 keY lel5 cm-2 B+ implant in a Si substrate pre-amorphi sed with 30 keY Ge+ (amorphous thickness = 50 nm), the Boron concentration at the alc interface after the implant is as low as Ixl017 cm-3. Figure I reports a typical result obtained in this case after annealing at 800°C, showing that Boron trapping at the EOR defects is most effective at the beginning of the anneal (when {311} defects are formed) and disappears at longer times, when {311} transform into {Ill} DLs. Similar behavior has been found in other works, indicating that Boron trapping at the EOR regions is most effective at temperatures around 800°C [8] and disappear at higher temperatures. It is important to note that, in such conditions, the EOR defect evolution is unaffected by the presence of Boron, i.e. defects nature, density and size distributions are found to be equal to those observed in the absence of Boron.

A detailed investigation of the efficiency of the EOR defects in trapping Boron was achieved by using dedicated test structures [9] such as the one depicted in Figure 2a, where two uniformly Boron-doped wafers

..,

c ... -t­Q. ! -t-� -t­c

Trapped B peak

:ti -t- 1-----\ -a -t-.. e --+

-«

..,

1 -'"

3x10'8 <0 11> a.. c: e 0 III 2xl0,8

lxl0'S

2.8x1020

(a)

- l>- 700'C_Exp - 0- 800'C_Exp - 0- 900'C_Exp

0--- -...0" ,

(b)

, ,

,_ .A - - , ..1>-

Depth

, ,

, ,

, ,

f

'I:

� 2.6xl020 -'" n:l 11> a.. 2.4xl020 c e S 2.2xl020

10 100 1000 Anneal Time [s]

Figure 2 - (a) Schematic illustration of the experimental

setup for the investigation of Boron trapping in B

uniformly-doped substrates. (b) Evolution of the Boron

concentration at the trapping peak for the wafer uniformly

doped with 2xl018 B/cm2. (c) Evolution of the Boron

concentration at the trapping peak for the wafer uniformly

doped with Ix1020 B/cm2.

(2xl018 and lx1020 cm-3, respectively) were amorphised with 30 keY Ge+ to a depth of 50 nm. Figure 2b reports the results from the low-concentration B-doped wafer, where the Boron concentration at the trapping peak (measured by SIMS) is reported as a function of annealing time for different temperatures. These results confirmed that boron trapping is mainly driven by the evolution of {31I} defects and that the {1I1} D Ls contribution to the trapping mechanism is less pronounced. In addition, an analytical model for the concurrent boron trapping (in the form of neutral BT pairs) at {311} defects and

--RT lE21 --600C/1h

- -- 800C/1h "e --900C/1h

u -- 900C/5h :;, III 'E' .2 lE20 ni ... -c CD u C 0 U

-CD lE19

o 50 100 150 200 250 300

Depth (nm) Figure 3 - Boron concentration profiles during annealing at

different thermal budgets. B+ implant: 10 keY, SXlO15 cm-2•

{Ill} DLs was developed [9] which takes into account the geometry of the EOR defects and has been implemented in the commercial SProcess TCAD simulator [10] since 2009.

A completely different behavior of the Boron concentration at the trapping peak was found in the highly-doped wafer (Ixl020 B/cm3), as shown in Figure 2c, where the peak magnitude constantly increases with annealing time for all temperatures. In addition, only a small concentration of defects was revealed by TEM analysis, in agreement with previous reports on the impact of B doping on {311} defect formation [II]. In such case, the concentration of Boron largely exceeds its equilibrium solid solubility limit at the investigated annealing temperatures. The Boron trapping peak is therefore associated with Boron precipitation in the form of Boron-Interstitial Clusters which are known to be harder to dissolve than {311} defects.

The "survival" of such BICs-driven Boron trapping peak at higher temperatures depends on the possibility of BTCs to grow up to large and stable defects. Whether or not BICs can reach such stable configurations, has been the subject of several studies. Before presenting further results on the Boron trapping in the presence of Fluorine (sections 4 and 5) we will therefore present in next section a recent investigation on the precipitation of Boron in heavily-doped silicon.

3. Boron precipitation in USJs: large clusters

formation

Experimental evidence of "large and stable clusters" in the presence of high concentration of implanted boron was provided by TEM investigations in the late 1990s [12] in the case of B+ amorphising implants (up lxlO17 B/cm2). Other TEM investigations

Figure 4 - Cross-section TEM micrographs from -implanted

silicon samples (10 keY, SxI015 cm-2) after annealing at

different temperatures and times. (a) 600-C/I h. (b) 800"C/I h.

(c) 900-Cllh. (d) 900-CISh.

were carried out after non-amorphising "USJ-relevant" boron implants (0.5 keY, Ixl015 B/cm-2) [13,14] which revealed the presence of dislocation loops lying on or close to (001) planes ((OOl)DLs), and being large enough to contain several hundreds of atoms. In addition to the habit plane, the (001 )DLs differ from the typical EOR defects with respect to their Burgers vector, which was found to be oriented either on the [00 I] direction or along the 4 <0 I I > directions lying out of the (001) plane. In all these studies, the exact chemical composition of the observed clusters could not be determined. In parallel, the recent development of the atom probe tomography (APT) technique [15] has made possible to measure the chemical composItIOn of B-Si precIpItates in boron supersaturated systems [16,17], showing that at low temperatures «800°C) the boron concentration in the observed precipitates remains well below the expected value in "stable" boride phases (SiB3, SiB4 [18]). However, no direct evidence has yet been given that the various defects observed using different techniques (TEM, APT) actually correspond to the same objects. In addition, it is not clear whether boron precipitates formed after high-dose non-amorphising boron implants can evolve towards stable phases.

In a recent study [19], the experimental conditions were therefore carefully chosen in order to investigate the Boron precipitation phenomenon by SIMS, TEM and ATP techniques. Boron ions were implanted at 10 keY to a dose of 5x1015 cm-2. Samples were then annealed at 600°C for Ih, 800°C (Ih) and 900°C (Ih and 5h).

Figure 3 shows SIMS profiles from all samples. The maximum boron concentration (�8x1020 cm-3)

1,0 60

:§!. 50 � Q)

40 � .� 30 a. .s

20 g o o

10 e 0,4 -'-----r---.,-----r-------,----1 0

o III

Figure 5 - Total Boron fraction in the precipitates (red squares,

left axis) and average boron atomic concentration in the

precipitates (blue circles, right axis) in B+-implanted silicon

samples (10 keY, 5xlOl5 cm-2) after annealing at different

temperatures and times.

exceeds the boron solubility limit in silicon for all the temperatures. Boron is therefore expected to precipitate into immobile clusters. This is clearly visible at high temperature 900°C, where Boron exhibits an immobile peak at the projected range depth.

TEM analysis, presented in Figure 4, indicates that at low temperatures (600°C, 800°C), the majority of the observed defects consists of "conventional" EOR defects ({311}s and {111}DLs), while at higher temperatures (cf. Figure 4c,d) large (001 )DLs are clearly observed, as previously found in B-rich silicon.

Finally, 3D boron reconstruction maps obtained by APT analysis reveal the presence of boron-enriched zones at depths corresponding to the implanted profiles. Combining top and side views of the precipitates it results that they have a circular platelet shape, and lye parallel to the (00 I) surface plane, therefore corresponding to the (00 I )DLs observed by TEM. Tn addition, the Boron concentration in the precipitates is found to continuously increase with increasing thermal budget (cf. Figure 5, circles, right axis) up to an average value of 50 at.% at 900°C 5h, with isolated clusters reaching the stoichiometric SiB3 composition (75 at.%).

[t is important to note that previous investigations carried out after low energy B implants (0.5 keY ) indicated that (001 )DLs are not stable at high annealing temperature [20]. When compared to them, the results presented here therefore demonstrate that the precipitation induced by high-dose boron implantation can indeed lead to the formation of the expected phase as given by the phase diagram (SiB3), provided the Boron and interstitial supersaturations created by the implant are sufficiently high and far from the silicon surface to avoid (or retard) surface

..--.. '" 'E � 1020 C o � 1019 +-' C

� 1018 c o u

- As implanted -800'C, 120 sec __ 800'C, 300 sec -- 800'C, 2700 sec

EOR defects

CO 1017 (a) ������ - as i m planted

c o :;::; � 1019 C Q) U C o U 18

u..

-- SOO'C, 120 sec -- SOO'C, 2700 sec

10 (b) �--�--�--��--�--�

o 20 40 60 80 100 Deoth (nm)

Figure 6 - (a) Boron (a) and Fluorine (b) concentration profiles

in Ge+-pre-amorphised USJs, after co-implantation with

Fluorine at 22 keY and annealing at 800°e. B+ implant

conditions: 0.5 keY, Ixl015 cm_2

interstitial recombination.

4. Effect of Fluorine on EOR defect formation and

Boron trapping

Before discussing Boron trapping experiments in BFiBF3 implanted junctions (next section), where high concentrations of both Boron and Fluorine are present below the alc interface after the amorphising implant, we will recall in this section the main [mdings concerning the simpler case where only Fluorine is present at high concentration in the EOR region, while Boron concentration remains low [3].

The conditions used for this experiment were similar to those reported in section 2 (0.5 keY lel5 cm-2 B+ implant in a Si substrate pre-amorphised to a depth of 50 nm). The main difference being an additional Fluorine implant (22 keY lxl015 cm-2), resulting in (i) an increased thickness of the amorphous thickness (60 nm) and (ii) a F concentration of about lxl020 cm-3 below the alc interface (cf. Figure 6b, squares). Tn comparison, the Boron concentration at 50 nm depth is as low as Ixl017 cm-3.

Figure 7 reports typical TEM micrographs

Figure 7 - TEM micrographs of B USJs preamorphized with

Ge after annealing at SOO°C for 300s. (a) Without additional p+ implant. (b) With additional F+ implant (22 keY, 1015 cm-2)

obtained after annealing at 800°C for 300s, without (a) and with (b) the additional F+ implant, showing that the defect population is drastically modified after the Fluorine implant: not only the apparent defect size and density are clearly larger, but it almost entirely consists of dislocation loops compared to the "B only" case (�70%). The systematic investigation of several annealing conditions allowed to conclude that in addition to the increased concentration of ex�ess interstitials (by a factor of � 10), the main impact of a Fluorine implant on the EOR defect evolution is the acceleration of the kinetics of the transformation of {311}s into dislocation loops, therefore leading to a stabilization of the defect population.

Considering that direct B-F interactions can be excluded [21,22], the effect of Fluorine on the Boron trapping is expected to be indirect, i.e. only due to the modified defect population. The results presented in Figure 6a show that B trapping at the EOR defects is slightly more effective in the presence of Fluorine, compared to the reference case (i.e. no Fluorine). Indeed, a Boron trapping peak is observed at 800°C even after long annealing times (300s and, to a lesser extent, 2700s), where B trapping was no longer detected in the absence of Fluorine after 300 s [23] (or 1000 s, cf Figure 1). This behavior can be explained

1020

1019

BF3

1018 0

- 0 - BF2 (beamline) as-implanted -.- BF2 (beamline) annealed

BF3 (plasma) as-implanted

-A- BF3 (plasma) annealed

�BF2

10 20 30 Depth (nm)

40

Figure S - Boron concentration profiles in USJs obtained with

BF2 beamline (4.5 keY, squares) or BP3 plasma immersion

implants (4.5 kY, triangles), after spike anneal at 1050°C.

considering that Fluorine induces an accelerated transformation of {311} defects (highly effective in trapping BI pairs) into less trapping-effective dislocation loops. However, this effect is counterbalanced by the large increase of dislocation loop size and density induced by Fluorine, which results in a much higher number of trapping sites in the EOR defects region.

5. Boron trapping in BF2/BF3 amorphised USJs

. Current USJs fabrication technologies routinely Involve the use of BF2 beam line ion implants (12) for both extension and SID contact fabrication. In parallel, advanced implant technologies such as Plasma !mme�sion ion implantation (PITT) are also being InvestIgated to achieve USJs with increased throughput characteristics, where BF3 molecules are �ypically used. In both cases (f and PIll), the high Implant doses typically used (of the order of lx1015 cm-2) combined with the low implant energy «S keY ) result in the amorphisation of a shallow surface layer «10 nm), with Boron and Fluorine concentrations below the alc interface as high as lxl020 cm-3. In such conditions, the combined effects of high B and F concentration and of surface proximity make it difficult to predict whether Boron trapping at the EOR defects can occur in these extreme conditions and how it would evolve during annealing.

Figure 8 shows B SIMS profiles taken before and after anneal (spike 10SO°e) following either a beamline BF2 implant (4.S keY, IxlO15 cm-2, green squares) or a BF3 plasma ion implant (4.S kV, lxl015 cm-2). Boron trapping peaks are observed at depths corresponding to the respective alc interface positions (cf. dashed lines in Figure 8), suggesting that, unlike

...--- 1021

'? E u

'-"

c 1020 0 :;:::; l\I L.. -C (J) 1019 U C 0

u 1018

0 20 40

- F as implanted

° B as implanted

• B BOO·C-10min

• B 900·C-10min

• B 1065·C-1sec

60 80 Depth (nm)

100

Figure 9 - Boron (symbols) and Fluorine (line) concentration

profiles in silicon implanted with BF3 (plasma immersion) at

10 kV (5x lOIS cm-2) after different anneals.

the experiments discussed in previous sections, the defects responsible for the observed trapping might be stable even at such a high annealing temperature. However, TEM analysis of such structures indicates that no extended defects are observed after the BF 3 plasma implant (not shown), while the residual defects visible after the BFz implant are too small to allow a reliable identification of their nature, probably due to the proximity of the surface.

In order to facilitate the structural investigation of the EOR defects in such experimental conditions, a dedicated experiment has therefore been designed where both the implant dose and energy of a BF3 plasma implant have been increased (10 kV, 5x1015

cm-2) with respect to the typical technology relevant conditions, so that the defects can evolve to large enough sizes for a reliable analysis. Figure 9 shows Boron SIMS profiles (symbols) taken before and after anneal at various temperatures (800°C 10 min, 900°C 10 min, 1065°C Is). In addition, the Fluorine concentration as-implanted profile is also shown (solid line). The Boron and Fluorine concentrations at the alc interface depth after the implant (16 nm) are 1 x 1 020 cm-3 and 3x10z0 cm-3, respectively. Cross-section TEM analysis from all investigated samples (not shown) reveals that EOR defects are present at a depth of �20

nm, and exhibit a weak evolution in defect size while the apparent defect density appears to decrea;e with increasing thermal budget. As expected, Boron trapping at the EOR defects is therefore observed for all annealing conditions, including the highest temperature (cf. Figure 9).

A detailed analysis of the defects formed in the sample annealed at 800°C for 10 min is reported in Figure 10, where a series of plan view images taken with different diffracting vectors g ([400], [040] and [2 -20]) with both positive and negative values of the

Figure 10 - TEM micrographs from a silicon sample implanted

with BF3 (plasma immersion) at 10 kV (5x101S cm·2) after

annealing at 8000e 10 min with different diffracting vectors, g, and deviation parameter, s.

deviation parameter, s, is reported. From this analysis it is concluded that the vast majority of the defects formed in these structures consist of dislocation loops lying on or close to (001) planes ((OOl)DLs), while the Burgers vector is oriented along one of the 4 <011> directions lying out of the (001) plane (defects labeled 1 to 4 on Figure 10). These defects are identical to those discussed in section 3 and therefore correspond to large Boron precipitates formed below the original alc interface during anneal due to the high Boron concentration at this depth after the implant. They are clearly responsible for the observed Boron trapping in the SIMS profiles. On the other hand, only a small fraction of the defects consist of {Ill} dislocation loops (defect labeled "A" in Figure 10), and this despite the presence of a high concentration of Fluorine at the alc interface after the implant. In any case, due to their low density, {III} DLs are not expected to strongly contribute to the observed Boron trapping in BFrimplanted USJs.

6. Conclusions and Summary

We have presented a review of some recent experimental studies on the Boron precipitation and trapping in pre-amorphised USJs. These studies suggest that the physical mechanism governing the Boron trapping mainly depends on the Boron

concentration left below the alc interface after the implant: • When the Boron concentration is low «1018 cm-\

the EOR defect evolution is unaffected and occurs with the formation of both {311} defects and {Ill} DLs. In such case, boron trapping is mainly driven by the evolution of {311} defects, while the {Ill} DLs contribution is less pronounced. As a consequence, it is most effective at temperatures

around 800°C • For higher Boron concentration values (�lxI020

cm-3 and beyond), boron trapping occurs through the formation of large Boron precipitates (in the form of (001) loops) that are stable up to high temperatures (> 1000°C).

In addition to providing a contribution to the understanding of the Boron trapping phenomenon in pre-amorphised USJs, these results clearly indicate that physical models for the formation of large Boron precipitates need to be implemented in TCAD simulators for a comprehensive description of the USJ fabrication process.

Acknowledgments

This work is partially supported by the EU project ATEMOX (lCT Project 258547). The authors

are very grateful to CEMES-CNRS laboratory in Toulouse (France) and also to Max-Planck Institute in Dusseldorf (Germany) for the use of their TEM and A PT facilities.

References

[1]. B . .T. Pawlak, R. Surdeanu, B. Colombeau, A..T. Smith, N.E.B. Cowern, R. Lindsay,W. Vandervorst, B. Brijs, O. Richard, F. Cristiano, App/. Phys. Lett. 84 (2004) 2055

[2]. F.Khaja, B.Colombeau, T.Thanigaivelan, D.Ramappa, T.Hemy, Proc. lITlO, edJ.Matsuo et a/., AlP 2010, p.65

[3]. S. Boninelli, F. Cristiano, W. Lerch, S. Paul and N. E. B. Cowern, Electrochem. Sol. State Lett., 10 (2007) H264

[4] . .T.H.Kim, L.M.Rubin, .T.YYoon, M.S.Ameen, T.S.Jang, .T.C.Cha, YH.Joo, A.B.Lee, S.WJin, Proc. JJTlO, edJ.Matsuo et al., ATP 2010, p.57

[5]. F.Gonzatti, F.Milesi. VDelaye, J.Duchaine. F. Torregrosa, H.Etienne, K. Y ckache, Proc. lITl 0,

edJ.Matsuo et al., ATP 2010, p.27 [6]. R. Duffy, Vc. Venezia, A. Heringa, T.WT. Husken,

M . .T.P. Hopstaken, N.E.B. Cowem, P.B. Griffin, C.C.Wang, App/. Phys. Lett. 82 (2003) 3647

[7]. S. Duguay, T. Philippe, F. Cristiano, D. Blavette, Appl.

Phys. Lett., 97 (2010) 242104 [8]. .1. J. Hamilton, N. E. B. Cowern, J. A. sSharp, K. J.

Kirkby, E . .T. H. Collart, M. Bersani , D. Giubertoni, A. Parisini, Appl. Phys. Lett., 89 (2006) 042111

[9]. E.M.Bazizi, P.F.Fazzini, C.Zechner, A.Tsibizov, H.Kheyrandish, A Pakfar, L.Ciampolini, C.Tavernier, F.Cristiano, Mat. Sci. Eng. B, 154-155 (2008) 275

[10]. http://www.synopsys.com/Tools/TCAD/ProcessSimulat IOn

[11]. T. E. Haynes, D . .T. Eaglesham, P. A. Stolk, H.-.T. Gossmann, D. C. Jacobson, and .T. M. Poate, Appl. Phys.

Lett., 69 (1996) 1376 [12].A. Agarwal, H.-J. Gossmann, D. J. Eaglesham, L. Pelaz,

S. B. Herner, D. C. Jacobson, T. E. Haynes, and R. Simonton, Mater. Sci Semicond Froc., 1 (1998) 17

[13]. F. Cristiano, X. Hebras, N. Cherkashin, A. Claverie, W Lerch and S. Paul. 2003, Appl. Phys. Lett., 83, 5407

[14].N.Cherkashin, M.Hytch, F.Cristiano, A.Claverie., Solid

State Phenomena, 108-109 (2005) 303-305 [15].B. Gault, F. Vurpillot, A. Vella, M. Gilbert, A. Menand,

D. Blavette, and B. Deconihout, Rev. Sci. lnstrum., 77 (2006). 043705-1-043705-8

[16]. O. Cojocaru-Miredin, D. Mangelinck, and D. Blavette, J. Appl. Phys., 106 (2009) 113525

[17]. M. Ngamo, S. Duguay, F. Cristiano, K. Daoud-Ketata, P. Pareige, J. App/. Phys., 105 (2009) 104904

[18]. T.B. Massalski. 1996, Binary alloy phase diagrams, 2nd edition

[19]. O. Cojocaru-Miredin, F. Cristiano, P-F. Fazzini, D. Mangelinck and D. Blavette, to be published

[20]. F. Cristiano, Y Lamrani, F. Severac, M. Gavelle, S. Boninelli, N. Cherkashin, O. Marcelot, A. Claverie, W. Lerch and S. Paul, Nuc!. lnstr. and Meth. B, 253 (2006) 68-79

[21].G. Impellizzeri, S. Mirabella, F. Priolo, E. Napolitani, and A. Camera, J. Appl. Phys. 99 (2006) 103510

[22].D.DeSalvador, G. Bisognin, E. Napolitani, M. Mastromatteo, N. Baggio, A. Camera, F. Boscherini, G. Tmpellizzeri, S. Mirabella, S. Boninelli, F. Priolo and F. Cristiano, App/. Phys. Lett., 95 (2009) 101908

[23].F. Cristiano, N. Cherkashin, P. Calvo, Y Lamrani, X . Hebras, A. Claverie, W. Lerch, S. Paul, Mater. Sci. Eng.

B 114-115 (2004) 174