From Solid-State Dewetting of Ultrathin, Compressively Strained

Mar 17, 2016 - From Solid-State Dewetting of Ultrathin, Compressively Strained. Silicon−Germanium-on-Insulator Films to Mastering the. Stoichiometry...
0 downloads 0 Views 2MB Size
Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

From Solid-State Dewetting of Ultrathin cSGOI Films to Si Ge Nanocrystals Stoïchiometry Mastering 1-x

x

Yann Almadori, Lukasz Borowik, Nicolas Chevalier, Wael Hourani, Frederique Glowacki, and Jean-Charles Barbe J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.6b01093 • Publication Date (Web): 17 Mar 2016 Downloaded from http://pubs.acs.org on March 20, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry C is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

From Solid-State Dewetting of Ultrathin cSGOI Films to Si1-xGex Nanocrystals Stoïchiometry Mastering

Yann Almadoria, Łukasz Borowik*a, Nicolas Chevaliera, Wael Hourania, Frédérique Glowackia and Jean-Charles Barbéa. a

Univ. Grenoble Alpes, F-38000 Grenoble, France

CEA, LETI, MINATEC Campus, F-38054 Grenoble, France

ABSTRACT

The mastering of material composition at nanoscale is of prime importance for many applications. In this context, we developed a simply implementable and fast germanium enrichment method allowing the control of the germanium fraction within silicon-germanium nanocrystals. In our process, the tuning of the stoichiometry is achieved in the background of thermally induced solid-state dewetting, followed by a solid-solid interfacial reaction step between nanocrystals issued from dewetting and the silicon-dioxide substrate. At first, we show by scanning transmission electron microscopy and scanning Auger microscopy that homogeneous silicon germanium nanocrystals are successfully formed by means of silicongermanium-on-insulator ultra-thin films solid-state dewetting. Then, we demonstrate that an 1 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 37

interfacial reaction between the silicon-germanium of nanocrystals and the silicon dioxide of the substrate can be thermally induced under ultra-high vacuum conditions. The reaction implying almost exclusively silicon, we show that it permits to modify the germanium fraction within silicon-germanium nanocrystals controlling its advancement. Finally, we highlight the possibility to control the stoichiometry of silicon-germanium nanostructures simply by using thermal effects.

INTRODUCTION During the past few years nanocrystals attracted a lot of research attention1-4 since they answer to the actual miniaturization requirements of many technological applications3,5 coupled with the possibility to tune fundamental physical properties such as band gap, radiative lifetime of optical excitation or even the melting temperature, via confinement effects simply by varying their dimensions.2,3,6 Consequently, several approaches allowing to synthetize nanocrystals with controlled dimensions have been developed recently.1,7-12 One of them is the thermally induced solid-state dewetting of ultrathin films. This method offers the possibility to form nanocrystals on a dielectric layer by a high temperature process, making of solid-state dewetting a challenging bottom-up technique which can be used for manufacturing microelectronic and/or nanoelectronic devices,13,14 for example memory15 or light emissive devices.16,17 It was shown that the realization of nano-catalyst particle patterns for the carbon nanotubes synthesis18 is also accessible by this process. To understand the mechanism of dewetting going from thin films to nanocrystals, Danielson et al.19 developed a thermodynamic model explaining the whole dewetting phenomenology of (001)-oriented silicon-on-insulator (SOI) thin films in 5 distinct steps: (i) critical void formation, (ii) void edge thickening with the formation of a rim at dewetting fronts preferentially oriented along stable crystallographic directions,20-23 (iii) void edge breakdown, (iv) dewetting fingers 2 ACS Paragon Plus Environment

Page 3 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

formation and growth mainly along directions24 and (v) destabilization followed by dewetting fingers splitting to form 3D nanocrystals (in the following the dewetted nanostructures can also be indifferently called agglomerates, nanoparticles or nanodots). This nanocrystals formation process is experimentally illustrated on figure 1.a displaying an AFM topography image showing partially dewetted area of a 7.5 nm thick (001)-oriented SOI thin film. In addition, it is shown that high temperature annealing process performed under ultrahigh vacuum conditions (UHV) may induce interfacial reactivity25-28 following the solidphase Si(s) + SiO2(s) → 2 SiO(g) reaction. This solid-phase interfacial reaction, favored at high temperature, implicates the silicon of nanostructures and the silicon dioxide (SiO2) of the buried oxide (BOx) to form gaseous silicon monoxide. As represented on figure 1.b, the interfacial reaction is revealed by the decreasing of nanocrystal dimensions, i.e. silicon amount, and the formation of pits within the SiO2 layer around nanocrystals.

Figure 1. AFM images of dewetted areas from SOI films: (a) before interfacial reaction (b) after partial interfacial reaction. On figure 1.b, we can clearly observe the formation of pits around silicon nanocrystals due to silicon-silicon dioxide interfacial reaction.

For most of the actual and future technologies, it appears that achieving a fine tuning of the system physical properties (e.g. optic and optoelectronic properties) at nanoscale is a key point to answer to requirements of many potential applications. In the particular case of 3 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 37

nanocrystals, two main approaches allow modifying the physical properties: (i) the control of the nanocrystal dimensions and (ii) the mastering of nanocrystals composition. Obviously, adjusting both parameters together is feasible as well. In the background of solid-state dewetting, the control of the nanocrystal dimensions after annealing is widely discussed in the literature and several methods are already developed.9,20,29-32 However, most of them need an important initial film preparation, limiting the implementation of such approaches. Moreover, these techniques do not permit the synthesis of nanostructures with monodispersed dimensions. On the other hand, huge efforts still have to be done to control the composition and the homogeneity of material at nanoscale.1,33 In this context, silicon germanium (SiGe) alloys seem to be one of the most promising candidate materials to design systems with tunable properties by mastering composition within nanocrystals. Indeed, as the bandgap in this material is directly proportional to the germanium relative concentration with respect to silicon one,34-37 their properties can be easily controlled to fit the requirements of many nanoelectronic applications38-40 by adjusting the germanium fraction. Moreover, silicongermanium-on-insulator (SGOI) ultrathin films, raw material to form silicon germanium nanocrystals by solid-state dewetting, are quite easily accessible since a large effort is currently done on the fabrication of this kind of films.38-41 In this work we developed a method to master the composition of SiGe nanostructures to take advantage of the peculiar properties of the two different systems described above, i.e. SiGe alloys and nanocrystals. We present a study of the thermally induced dewetting process of compressively strained silicon-germanium-alloy-on-insulator (cSGOI) ultrathin films, followed by a high temperature annealing process under UHV conditions in order to master the germanium fraction within the SiGe nanocrystals formed by dewetting. Our innovative germanium enrichment method is based at the same time on two distinct physical phenomena: 4 ACS Paragon Plus Environment

Page 5 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(i) the SiO2/Si interfacial reaction shown by several studies25-28 in case of silicon nanostructures deposited on SiO2 and (ii) the selectivity of the silicon atoms oxidation in SiGe alloys discussed by Tezuka et al.42 Accordingly, we expect an increase of the germanium fraction, i.e. germanium concentration, within SiGe nanocrystals as the SiO2/Si interfacial reaction between the nanocrystals and the BOx is occurring. Experimentally, our assumption should be illustrated by an increase of the germanium concentration as the nanocrystal dimensions become smaller. In this paper, we first focus on the solid-state dewetting phenomenology and mechanisms of cSGOI thin films, since to our knowledge no publications deal with this particular case. Then we show that solid-state dewetting of SGOI thin film efficiently leads to the formation of homogeneous SiGe nanocrystals. Finally, we demonstrate that the modification of the germanium fraction within SiGe nanocrystals can be achieved in UHV conditions simply adjusting the annealing process temperature during or following dewetting step.25-28,42

EXPERIMENTAL AND METHODS

Samples: The samples examined during this study consist on thin 12 ± 1 nm thick monocrystalline silicon germanium (Si1-xGex) alloy layers (001)-oriented, lying on a 147 ± 5 nm thick amorphous thermal SiO2 layer. The SGOI alloy thin films are obtained by the enrichment method41 giving rise to bi-axially strained SGOI. The bi-axially strain is estimated at 1.1 ± 0.1 GPa by mean of Raman spectroscopy.41 In this study, the thickness and the elemental composition of the Si1-xGex top layer are measured by means of x-ray reflectometry, x-ray fluorescence and spectroscopic ellipsometry as described by Vincent et al.43 In our samples case, the elemental composition is found to be of 22 ± 2 atomic percent (at. %) of germanium giving a film stoichiometry of Si0.78Ge0.22. A thin native oxide layer of

5 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 37

0.8 ± 0.2 nm is measured on the SGOI surface. In the paper, the compressively strained Si0.78Ge0.22 film/SiO2/Si substrate is called cSGOI systems. No particular treatment is done on our samples in order to prevent the impact of chemical treatment on the dewetting phenomenology. According to our previous study,44 we assume that surface contamination, particularly atmospheric carbon contamination, has a fairly low impact on the dewetting process if no chemical treatment is performed.

Solid-State Dewetting and annealing: The solid-state dewetting of square samples (10 × 10 mm²) of Si0.78Ge0.22 ultra-thin film is thermally induced under UHV conditions (1×10-10 mbar) using a preparation chamber from Omicron system. The solid-state dewetting of the cSGOI thin film is obtained by heating the sample at a temperature of 750 ± 10°C during about 15 minutes. The temperature is then progressively increased to reach 880 ± 10°C (with

ΔT ~ 10 °C.min-1) which are maintained during 30 minutes. This temperature is theoretically sufficient to induce interfacial reactions between Si nanocrystals and the BOx in the case of SOI dewetting, according to the literature.25-28 It is then obvious that this processing temperature is also sufficient in the case of cSGOI systems since the Si-Ge and Ge-Ge binding energies are weaker than the Si-Si one.45 The temperature is controlled by using a direct top surface temperature measurement with an infrared pyrometer with an accuracy of ± 10°C.

Characterization: The samples characterization is done after the sample is cooled down to room temperature and took out from the UHV sample preparation system. The size of the samples (10×10 mm2) is large enough to avoid influence of sample edges. Moreover, each characterization is realized at the center of the samples far from dewetting fronts.

6 ACS Paragon Plus Environment

Page 7 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

The topography of the dewetted surface, i.e. the morphology of dewetted nanocrystals or agglomerates, is observed at ambient atmosphere by atomic force microscopy (AFM)46 in Tapping mode using Dimension FastScan from Bruker. In this study, AFM images are acquired with FastScan-A cantilevers from Bruker and with 512*512 pixels. The Auger electron spectroscopy (AES) and scanning Auger electron microscopy (SAM) experiments are carried out with a PHI 700Xi Auger nanoprobe equipped with a cylindrical mirror analyzer (CMA). The CMA is mounted coaxially with the electron column with an energy resolution of 0.5% and high sensitivity (ultimate detection limit ∼0.1%). All the measurements are performed with an incident electron beam of 20 keV in energy and a 1 nA incident current. Under these conditions, the spatial resolution is about 20 nm. Quantitative Auger information is obtained from the Auger KLL transition peak intensities normalized by the average matrix relative sensitivity factors47 (AMRSFs) available in the MultiPak software v.9.5.0 of Physical Electronics. The scanning transmission electron microscope (STEM) experiments are realized on a FEI Tecnai Osiris microscope operated at 200 keV. The high angle annular dark field (HAADF) images are acquired in STEM mode. Energy dispersive X-ray (EDX) experiments are performed in STEM mode using the four windowless silicon drift detectors (SDD) integrated into the pole piece that allow the X-ray detection over a 0.8 srad solid angle. EDX maps are acquired with a probe current of ~ 300 pA over areas of 120×120 nm2, using pixel sizes of 0.16 nm, dwell times of 67 µs per pixel and an overall time of about 30 minutes with the drift correction activated. In our experiments, the TEM samples consist in thin lamella obtained by a focus ion beam (FIB) process. Elemental quantification of the experimental EDX spectra has been performed on the intensity of Kα Si and Ge peaks and using the Cliff–Lorimer factors at 200 kV available in the Esprit 1.9 Bruker software. Other elements appearing in the

7 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 37

EDX spectra (C, O, Cr, W, Cu) have been used for the peak deconvolution (based on Gaussian functions).

RESULTS AND DISCUSSION Figures 2.a and 2.b present AFM images of a partially dewetted area of a compressively strained (001)-oriented Si0.78Ge0.22 thin film at different steps of the dewetting process. We can directly point out an atypical dewetting phenomenology as compared with well-known dewetting mechanism for SOI films illustrated on figure 1.a. This atypical dewetting is also widely different from the one expected for strain-free SGOI thin films.48 Indeed, no long range periodical structures and no preferential directions are pointed out. The topology of the dewetting fronts is no longer correlated, in a long distance range, to the in-plane crystallographic orientations of the thin film and exhibit a very irregular shape. The appearance of such instabilities in the dewetting front has already been observed in the literature31,48 and described as secondary edge or short-wave length (SWL) instabilities. We can also note the absence of elevated rim at dewetting void edges. Moreover, nanocrystals seem to be formed directly from the dewetting fronts and typical dewetting fingers are not observed, suggesting that nanocrystals are not issued from the destabilization of dewetting fingers because of Rayleigh like instabilities.19 This last feature implies a random-like spatial distribution of the agglomerates on the BOx layer, as we can see on the fully dewetted area displayed on figure 2.c. The height histograms in figure 2.d highlight a bimodal diameter distribution with particles having a mean diameter close to 260 nm while other ones have a mean diameter of about 90 nm. The agglomerates surface density is measured to be 7.5 µm-², which is slightly higher than the 0.5 - 5 µm-² surface density expected for 12 nm thick layer of SOI.24,30,49 This higher density is then consistent with an instability length diminution.

8 ACS Paragon Plus Environment

Page 9 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 2. (a-c) AFM images of several dewetted areas of a cSGOI thin film taken at different steps of the dewetting process, respectively: (a) At the voids nucleation, (b) at an intermediate state and (c) after the sample is fully dewetted. (d) Diameter distribution of the agglomerates (associated to figure 2.c).

Based on our sample characteristics regarding to SOI and referring to the literature, three main hypotheses leading each to different effects can explain our present observations: (i) the presence of germanium within the SiGe ultra-thin film, (ii) the influence of strain as we worked with compressively–strained samples or (iii) a strain effect driven by surface contamination. According to Capellini et al.,48 which are to our knowledge the only ones to consider thermally induced solid-state dewetting of ultrathin SGOI films, the first hypothesis seems not to be satisfying. Indeed, they did not point out a strong influence of germanium presence within thin film during strain-free (001)-oriented Si0.8Ge0.2-OI films solid-state dewetting 9 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 37

((001)-oriented cSi0.78Ge0.22-OI in this study). For instance dewetting fingers, elevated rims and preferential orientations are still observed in their study. Another study50 focusing on the stability of ultrathin Ge/Si(001) films pointed out an influence of the thickness of the Ge layer on the dewetting mechanisms. However, they conclude that the impact on solid-state dewetting is due to the strain locally induced in the SOI film by the diffusion of germanium. Furthermore, several studies have shown that strain within thin SOI films leads to strong modifications of solid-state dewetting mechanisms and phenomenology.17,32,48 The main feature expected in that case is the presence of short-wave length instabilities along the dewetting fronts, while in some particular cases the long range ordering is also lost.50 The shorter period for edge instabilities is in good agreement with the surface density rising and the loss of long-range arrangement of nanocrystals observed in our study. However, strain seems not enough to explain the elevated rims and dewetting fingers absence in cSGOI case. Finally, the most probable hypothesis appears to be the third one. Indeed, Borowik et al.31 pointed out that the coupling of strain effects and the presence of a high initial carbon contamination of the upper surface of SOI thin films can lead to modifications matching pretty well with most of our experimental observations. We then assume that in the case of compressively strained SGOI systems, the initial carbon contamination has a pretty high influence on solid-state dewetting. In summary, according to the comparison of our results and the literature, it seems that a couple effects of strain and surface contamination is privileged to explain our experimental observations. However, it is difficult to clearly and unambiguously identify the reason of these observations, as several hypotheses of the literature can be satisfying and since a lot of parameters can lead to similar observations even for small modifications. The influence of such-and-such parameters still has to be fixed and will be the object of future studies. 10 ACS Paragon Plus Environment

Page 11 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

As already mentioned, the possibility to control the enrichment within SiGe nanocrystal is of prime importance for several applicative domains. Before considering the possibility of germanium enrichment at nanoscale, let focus on the characterization of the Si and Ge distributions within as-dewetted nanocrystals volume directly after their formation, i.e. before pits formation in the BOx layer. Figure 3.a is a cross-section HAADF-STEM image of representative as-dewetted agglomerate. Corresponding EDX mappings for germanium and silicon are respectively presented on figure 3.b and 3.c.

11 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 37

Figure 3. (a) HAADF-STEM image of as-dewetted agglomerate and the BOx. Elemental EDX maps associated to agglomerate shown in figure 3.a for (b) germanium and (c) silicon.

(d) In-depth AES measurements of the germanium fraction (atomic %) of raw cSGOI thin film (black filled circles) and partially dewetted cSGOI film near a dewetting front (orange empty squares) in function of the sputtering time, i.e. the depth. The red line curve, associated to the top axis, corresponds to the germanium concentration profile obtained by STEM-EDX measurements and displayed on figure 3.a (red dash-line). The Ge concentration seems to be higher inside as-dewetted agglomerates than in raw film.

First of all, we both observe, respectively on figures 3.b and 3.c, the presence of silicon and germanium within the agglomerates. Thus it appears that, by solid-state dewetting of SGOI thin films, we form nanocrystals almost exclusively composed of silicon and germanium, as we could expect. Moreover, referring once again to figures 3.b and 3.c, the both species (Si and Ge) distributions seem to be qualitatively homogeneous in the whole nanocrystal volume. Indeed, no Si or Ge rich domains are observed by STEM-EDX measurements although the spatial resolution is 0.16 nm. This result is highly different from what is observed in the case of SiGe nanocrystal obtained during heteroepitaxy growth process,1,33,51-54 where germanium rich and silicon rich domains are pointed out. Consequently, we demonstrate here that solidstate dewetting is a good candidate to produce homogeneously composed silicon-germanium nanostructures. In a more quantitative approach, we consider in figure 3.d, the comparison of in-depth germanium concentration profiles for: (1) as-received cSGOI thin film (raw Si0.78Ge0.22 film), (2) film close to a dewetting front and (3) SiGe agglomerate. The two first measurements (orange empty squares and black filled circles) are done by Auger spectroscopy, considering 12 ACS Paragon Plus Environment

Page 13 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Si and Ge KLL transitions,

while the last one (red full line) corresponds to the EDX

measurement. For nanocrystal measurements, EDX is used instead of AES since with Auger we cannot properly quantify the germanium concentration because of the BOx layer. Moreover an AES concentration Z-profile imposes to sputter the nanostructures. However, sputtering homogeneously such structures still remains a real experimental limitation. The profiles in figure 3.d reveal two main features. (i) First, the homogeneity of nanocrystals (also true for films), qualitatively observed previously, is confirmed since the germanium and silicon fractions are fairly constant from the extreme surface of the nano-dot to the BOx layer. Consequently, in the following we will consider that the surface information is representative of the entire dewetted nanostructure volume, making from Auger spectroscopy, which commonly gives information about the upper 3 nm of a sample, a relevant and powerful tool to probe the nanocrystals composition. (ii) Secondly, the germanium fraction is higher within nanocrystals than in the film. Indeed, the germanium fraction for the as-deposited cSGOI thin film is measured by Auger spectroscopy to be 21 ± 3%. This value is in fair agreement with the nominal value given at 22 ± 3% by initial measurements. Otherwise, according to EDX measurement, the relative amount of germanium surprisingly widely increases within dewetted nanocrystals since we obtain an almost constant concentration close to 35 ± 4% in the entire volume. Similar values are obtained for 3 distinct as-dewetted nanocrystals with height going from 90 nm to 120 nm making this trend quite robust. This means that the alloy stoichiometry passes from Si1-xGex with x = 0.22 for the thin film to Si1-yGey with y = 0.35 within nanocrystals. As there is no external Ge source in our experimental system, the clear Ge enrichment, observed between raw film and nanocrystals on figure 3.d, is consistent with the disappearance of a significant part of silicon material previously presents within the cSGOI film. According to the preferential and almost exclusive reaction of silicon during SiGe alloys 13 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 37

oxidation,41,42 an oxidation process is favored to explain the vanishing of silicon while germanium material seems not to be affected. As the dewetting is performed under UHV condition (P ~ 10-10 mbar), we consider that there is no significant active silicon surface oxidation according to the low oxygen amount in the sample environment during experiments. Under these experimental conditions only two oxygen sources can be identified: (i) the native oxide on the upper surface of the film and (ii) the BOx layer. Consequently, the silicon oxidation leading to the germanium enrichment can only be addressed to: (i) film native oxide removal by reaction with the SiGe alloy and/or (ii) interfacial oxidation reaction of the SiGe material with the BOx layer.25-28 A rapid calculus based on material conservation has shown that in our systems going from x = 0.22 to y = 0.35 in germanium fraction requires an initial native oxide thickness of at least 6 nm. The native oxide in our case being thinner than 1 nm (cf. experimental and methods section), the oxygen amount within the native oxide is too low to explain germanium enrichment in such proportions. The only abundant source of oxygen is the BOx. We thus assume, according to the observation of small grooves in the BOx layer close to agglomerate borders in figure 3.a, that an interfacial reaction occurs between BOx and SiGe alloy concomitantly to the dewetting process. Additionally, the germanium concentration close to the dewetting front is measured to be about 22 ± 3 %. Considering the errors bars, the small difference in germanium fraction between raw film and film close to an edge seems not to be really significant and suggests no evolution of the germanium and silicon concentrations between both locations. Thus, the proximity of germanium fraction within the film close to a dewetting front with respect to the raw film one may suggest that the observed phenomenon is directly linked to the nanocrystals formation. This last remark strengthens the assumption of an interfacial reaction, since it is in good agreement with the Sudoh scenario where interfacial reaction initiate at triple line27,28 14 ACS Paragon Plus Environment

Page 15 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(nanocrystal, BOx and vaccum interface). Moreover, in the SGOI case as well, the interfacial reaction starts at the early instant of the nanocrystals formation, even at “low” temperature. Finally, we demonstrate that an interfacial reaction, well known for SOI systems,25-28 also occurs in the case of SiGe agglomerates on top of a SiO2 layer. In this particular case, to explain the germanium enrichment, we assume a reduction of the BOx by the silicon of the silicon germanium alloy accompanied by the rejection of germanium from the reaction area and by its condensation.41,42 The associated reaction can be schematically written as SiGe(s) + SiO2(s) → 2 SiO(g) + Ge(s). Additionnaly the homogeneous germanium distribution in the whole volume of nanocrystals is not surprising since, according to Tezuka et al.42, the germanium is supposed to diffuse very fast through nanocrystals due to its high mobility in silicon at high temperature. Thus, we consider a “low temperature” (750°C) initial interfacial reaction to be the dominant phenomenon in silicon vanishing, leading to the germanium enrichment observed within as-dewetted nanocrystals with respect to the raw film. Obviously the enrichment effect can be little enhanced by native oxide reaction. In the following, it is important to well differentiate the unexpected interfacial reaction occurring concomitantly with the formation of “as-dewetted” nanocrystals during the dewetting process (T = 750 ± 10°C) from the induced interfacial reaction willingly provoked with high temperature annealing (880 ± 10°C).

Figure 4.a displays a scanning electron microscope (SEM) image of an as-dewetted cSGOI system, i.e. before the high temperature annealing process realized to induce huge interfacial reaction. On this image we can clearly identify nanocrystals formed by the dewetting process as well as the bimodal diameter distribution already observed by AFM and displayed in figure 2.d. On the other hand, as no pits in the BOx are observed next to dewetted nanoparticles, we confirm that no huge interfacial reaction occurred yet. 15 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 37

16 ACS Paragon Plus Environment

Page 17 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 4. SEM images of fully dewetted areas (a) before (as-dewetted nanocrystals) and (b) after high temperature annealing at 880°C. Elemental (c) germanium and (e) silicon SAM maps (KLL Auger transitions) associated to the area displayed in figure 4.a (before annealing at 880°C). Elemental (d) germanium and (f) silicon SAM maps (KLL Auger transitions) associated to the area displayed in figure 4.b (after annealing at 880°C). (g) Germanium over silicon Auger intensity ratio IGe/ISi measured by AES on different locations corresponding to the film, nanocrystals (NCs), nanocrystals edge and the BOx (black full circles (bottom xaxis)) and on different agglomerates (red empty diamonds (top x-axis)) before annealing at 880°C. (h) Germanium over silicon Auger intensity ratio IGe/ISi measured by AES on two different areas (area-1 displayed by empty black diamonds (bottom x-axis)) and area-2 displayed by full red circles (top x-axis)) for agglomerates with increasing diameters after annealing at 880°C. The area-1 corresponds to the one shown on the SEM image displayed in inset. Area-2 is a similar zone measured in another sample (SEM image not shown) and annealed under the same conditions. As the diameter of the agglomerates increases (1 to 6), the IGe/ISi ratio increases as well, suggesting a higher germanium fraction within smaller nanoparticles.

Figures 4.c and 4.e display the elemental mappings of the area shown in figure 4.a (before annealing at 880°C). They are respectively associated to the contributions of germanium and silicon Auger KLL transitions. Once again, we can notice on figure 4.c the presence of germanium within as-dewetted nanocrystals. Moreover, it qualitatively seems that the signal is constant whatever the considered agglomerate and its dimensions. Thus, this result agrees with the previous EDX measurements (fig. 3.b) and confirms the formation of homogeneous SiGe nanocrystals. Moreover, it appears that almost all the germanium material is contained within agglomerates, according to the absence of germanium Auger contribution on the BOx 17 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 37

layer. The silicon Auger KLL transition mapping displayed on figure 4.e presents a low contrast between nano-dot centers and the BOx layer while a higher contrast is observed on agglomerate’s edges. In the case of such nanostructures, as the primary electron penetration depth is large with respect to the nanocrystal height, the electron beam may induce Auger transitions in the BOx via secondary electrons.55 These transitions can then be dominant in the total Auger signal and are able to explain the low contrast on the particle center compared to the BOx layer. This feature well known for silicon nanostructures deposited on SiO2 layer55 does not permit to properly quantify the fraction of silicon within the SiGe agglomerates. On the other hand, the higher contrast on the agglomerate edges is addressed to SAM edge effects55,56 coupled with the finite size of the electron beam nano-probe which also induces Auger transition in the BOx layer. Despite the fact that it is difficult to extract information from silicon Auger map (fig. 4.e), no size effects can be pointed out suggesting that the silicon distribution is homogeneous within as-dewetted nanocrystals. In summary, these qualitative observations seem to indicate that all the compounds are homogeneously spread over the nanocrystals surface and are by extension, and in agreement with STEM-EDX measurements (cf. fig. 3), homogeneously distributed in volume. In figure 4.g the intensity ratio IGe/ISi between germanium and silicon Auger KLL transitions is considered in order to get semi-quantitative informations about the fraction evolution of both species. The empty red diamonds are related to measurements done on the central part of five different as dewetted nanocrystals (cf. figure 4.a) while the full black circles correspond to different locations on the sample surface. The fairly constant IGe/ISi ratio measured on several as-dewetted nanocrystals (NCs in figure 4.g) indicates a homogeneous germanium concentration whatever the considered agglomerate and its dimensions, confirming the qualitative observation done on figure 4.b. Moreover, the very low IGe/ISi ratio of about 0.05 (almost 0) measured on the BOx corroborates that almost all the germanium is 18 ACS Paragon Plus Environment

Page 19 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

contained within SiGe nanocrystals. We also point out a difference in the IGe/ISi ratio obtained on agglomerates and on the SiGe thin film with values respectively of 0.8 and 0.6. This feature seems to confirm the germanium enrichment inside as-dewetted nanocrystals compared to the raw cSGOI thin film, as it was previously observed in figure 3.d. Furthermore, the proximity of IGe/ISi ratios, close to 0.8, between agglomerates (NCs) and agglomerate’s edges (NCs edges) is in line with an artefact of measurement (edge effect) rather than a real surface modification of the silicon amount (cf. fig. 4.e).

Figure 4.b presents a SEM image acquired in a fully dewetted area of a sample annealed at 880°C during 30 minutes in order to induce huge interfacial reaction between SiGe nanocrystals and the SiO2 BOx layer. In this micrograph, we can clearly observe the formation of holes/pits in the BOx. Some of them still contain SiGe nanocrystals while some others appear empty. In the same time, figure 4.b shows that nanocrystals clearly exhibit various diameters. It is then pointed out that most of the dewetted agglomerates have a lower mean diameter after the high temperature annealing process (fig. 4.b) compared to asdewetted agglomerates (fig. 4.a). This dimensions diminution clearly indicates a material loss. Moreover, it qualitatively seems that the nanocrystals diameter is linked to the surrounding hole dimensions. In other words the larger the pit, the smaller the nanocrystal. Thus, in agreement with the holes formation and the decrease of nanocrystal dimensions observed in figure 4.b, it appears that significant interfacial reaction occurred between BOx and nanocrystals. Experimentally, according to the SiGe(s) + SiO2(s) → 2 SiO(g) + Ge(s) reaction, these features might be revealed by a higher germanium fraction within small nanoparticles, where interfacial reaction is more advanced, with respect to bigger ones. In addition as already discussed, we expect an homogeneous germanium enrichment in the whole volume of

19 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 37

nanocrystals since, according to Tezuka et al.42, the high temperature mobility of germanium in silicon is important. Figures 4.d and 4.f display elemental germanium and silicon Auger KLL transition mappings of the area shown in figure 4.b. It clearly appears on figure 4.d that small agglomerates exhibit higher germanium Auger intensity than larger particles. Thus, this feature suggests that the germanium concentration is higher within the smaller nanocrystals. In return the silicon fraction seems lower in small particles as supposed by the impossibility to distinguish them in figure 4.e while we can do it for larger nanocrystals. In summary, when pits are formed around nanocrystals, the germanium fraction increases while the silicon one decreases as dewetted agglomerates become smaller. Thus the results presented so far suggest the possibility to modify the germanium fraction within nanostructures controlling their dimensions, in other words controlling the interfacial reaction process advancement. Figure 4.h display the intensity ratio IGe/ISi between the germanium and the silicon Auger KLL transitions, after the high temperature annealing process (880°C), for nanocrystals with decreasing dimensions (D1>D2>…>D6, where Di is the mean diameter of the particle i). In this figure, measurements associated to area-1 correspond to the black dots displayed on the inset of figure 4.h. Area-2 is a similar zone measured in another sample (not shown) and annealed under the same conditions. In this case, the modification of the nanocrystals dimensions, i.e. diameter, can be related to the interfacial reaction advancement. The smaller the nanocrystal, the more silicon reacted. First observations clearly indicate a strong nanocrystal dimensions dependence of the IGe/ISi ratio. Indeed, on figure 4.h the IGe/ISi ratio increases as the nanocrystals diameter decreases. In the area-1, for example, nanocrystals 1, 4 and 6 with diameters 190, 105 and 80 nm display IGe/ISi ratios of 0.8, 1.35 and 1.8, respectively. The same behavior is pointed out in the area-2 20 ACS Paragon Plus Environment

Page 21 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

with an IGe/ISi ratio variation with respect to the nanocrystals mean diameter, going from 0.80.9 to 2.1 for nanocrystals with diameters of 230 and 60 nm, respectively. Note that for the largest nanocrystals, i.e. where the reaction is the less advanced, the IGe/ISi ratio is close to 0.8, in good agreement with the value measured in the case of as-dewetted nanocrystals (cf. fig. 4.g). Moreover, the IGe/ISi ratio increases by a factor close to 2 in the considered nanocrystals range, regardless to the measurement area. Since the IGe/ISi ratio is directly related to the relative fraction between silicon and germanium inside silicon germanium alloys, it appears that the germanium fraction increases as the induced interfacial reaction is more and more advanced, i.e. when the nanocrystal dimensions decreases. We thus have demonstrated here the possibility to influence the germanium fraction within silicon germanium nanostructures, modifying the nanocrystals diameter, i.e. the interfacial reaction advancement.

As discussed previously in this paper, our enrichment process qualitatively well works. However some limitations still remain. First of all, the initial SiGe nanocrystals dimensions dispersion after solid-state dewetting is so far a strong difficulty. Indeed, one can easily understand that larger SiGe nanostructures contain a higher amount of silicon. Consequently, the process duration needed to obtain nanocrystals with a given germanium fraction will be dependent of the initial dimensions of the nanocrystals. In other words if we do not have a monodispersed distribution of nanocrystal dimensions it will probably not be possible to get a monodispersed distribution of germanium fraction within nanocrystals using a single annealing process. On the other hand, as we can see on the germanium mapping displayed in figure 4.d, in some pits the germanium is not detected at all by SAM measurements. Two main hypotheses permit to explain this absence of germanium signal. The most probable is that we are not able to detect small germanium nanocrystals located in the bottom of the holes 21 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 37

because of the narrowness of holes and the impossibility for Auger electron to exit from them. On another side, a reaction of the germanium with the BOx followed by germanium oxide desorption is not totally excluded after all the silicon has reacted. In this case, our process would not permit to form pure germanium nanocrystals. However, this hypothesis is not favored since reduction of SiO2 by germanium is energetically disadvantageous. Moreover in classical germanium enrichment technique it is possible to form pure germanium thin films.41,42

Finally, despite the fact that some limitations exist and that further studies are needed to optimize our process, we successfully developed and characterized an innovative germanium enrichment method under UHV conditions allowing, if totally controlled, the mastering of the germanium fraction in germanium alloys at nanoscale simply using thermal effects.

CONCLUSION In this paper, we highlighted the possibility to influence the germanium fraction at nanoscale in SiGe nanocrystals. This is illustrated in the background of cSGOI ultra-thin film solid-state dewetting, through solid-phase interfacial reaction of silicon germanium with SiO2. Results show an atypical dewetting process addressed to the coupled effect of strain and the influence of initial surface contamination of the ultrathin films. Otherwise chemical analysis of 3D nanocrystals formed during dewetting process revealed a quite homogeneous spreading of germanium and silicon materials within the entire volume of these nanostructures. The quantitative analysis of each component also shows that germanium concentration is higher in dewetted nano-agglomerates with respect to the nominal cSGOI thin film germanium concentration. This observation is thus associated to an initial interfacial reaction concomitant 22 ACS Paragon Plus Environment

Page 23 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

to the formation of nanocrystals during the solid-state dewetting process. Then based on the assumption of a preferential silicon and SiO2 interfacial reaction occurring between SiGe nanocrystals and the BOx layer, we demonstrated using a high temperature annealing process under UHV conditions that it is possible to tune the germanium concentration within nanocrystals smaller than 100 nm. Indeed, it appears that the germanium fraction can be adjusted controlling the agglomerates diameter, i.e. controlling the interfacial reaction advancement. This method is promising since it permits an easy and fast implementation as the main parameters to control are the annealing temperature and the process duration. It is thus of real interest since it opens ways for SiGe nanocrystals with tuned Ge concentration to innovative technological applications.

AUTHOR INFORMATION

Corresponding Author *E-mail: [email protected]

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work was supported by the French “Recherche Technologique de Base” Program and performed in the frame of the ANR LOTUS project. The measurements were realized on the CEA Minatec Nanocharacterization Platform (PFNC). We particularly thank N. Bernier for the STEM-EDX measurements.

23 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 37

REFERENCES (1) Aqua, J.-N.; Berbezier, I.; Favre, L.; Frisch, T.; Ronda, A. Growth and SelfOrganization of Sige Nanostructures. Physics Reports 2013, 522, 59-189. (2) Xia, Y.; Yang, P.; Sun, Y.; Wu, Y.; Mayers, B.; Gates, B.; Yin, Y.; Kim, F.; Yan, H. One-Dimensional Nanostructures: Synthesis, Characterization, and Applications. Adv. Mater.

2003, 15, 353-389. (3) Alivisatos, A. P. Semiconductor Clusters, Nanocrystals, and Quantum Dots. Science

1996, 271, 933--937. (4) Borowik, L.; Lepage, H.; Chevalier, N.; Mariolle, D.; Renault, O. Measuring the Lifetime of Silicon Nanocrystal Solar Cell Photo-Carriers by Using Kelvin Probe Force Microscopy and X-Ray Photoelectron Spectroscopy. Nanotechnology 2014, 25, 265703. (5) Schmid, G. Nanoparticles: From Theory to Application. WILEY-VCH: Weinheim, 2011. (6) Barbagiovanni, E. G.; Lockwood, D. J.; Simpson, P. J.; Goncharova, L. V. Quantum Confinement in Si and Ge Nanostructures. J. Appl. Phys. 2012, 111, 034307. (7) Chou, S. Y.; Krauss, P. R.; Renstrom, P. J. Nanoimprint Lithography. J. Vac. Sci. Technol. B 1996, 14, 4129-4133. (8) Gago, R.; Vàzquez, L.; Cuerno, R.; Varela, M.; Ballesteros, C.; Albella, J. M. Production of Ordered Silicon Nanocrystals by Low-Energy Ion Sputtering. Appl. Phys. Lett.

2001, 78, 3316-3318. (9) Ishikawa, Y.; Kumezawa, M.; Nuryadi, R.; Tabe, M. Effect of Patterning on Thermal Agglomeration of Ultrathin Silicon-on-Insulator Layer. Appl. Surf. Sci. 2002, 190, 11-15. 24 ACS Paragon Plus Environment

Page 25 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(10) Chen, A.; Chua, S. J.; Chen, P.; Chen, X. Y.; Jian, L. K. Fabrication of Sub-100 Nm Patterns in Sio2 Templates by Electron-Beam Lithography for the Growth of Periodic Iii--V Semiconductor Nanostructures. Nanotechnology 2006, 17, 3903. (11) Park, J.; An, K.; Hwang, Y.; Park, J.-G.; Noh, H.-J.; Kim, J.-Y.; Park, J.-H.; Hwang, N.-M.; Hyeon, T. Ultra-Large-Scale Syntheses of Monodisperse Nanocrystals. Nature materials 2004, 3, 891--895. (12) Pavesi, L.; Dal Negro, L.; Mazzoleni, C.; Franzo, G.; Priolo, F. Optical Gain in Silicon Nanocrystals. Nature 2000, 408, 440--444. (13) Liao, W. M.; Li, P. W.; Kuo, D. M. T.; Lai, W. T. Room-Temperature Transient Carrier Transport in Germanium Single-Hole/Electron Transistors. Appl. Phys. Lett. 2006, 88, 182109. (14) Wu, J. H.; Li, P. W. Ge Nanocrystal Metal-Oxide-Semiconductor Transistors with Ge Nanocrystals Formed by Thermal Oxidation of Poly-Si0. 88ge0. 12. Semicond. Sci. Technol.

2007, 22, S89. (15) Chen, J.-H.; Lei, T.-F.; Landheer, D.; Wu, X.; Liu, J.; Chao, T.-S. Si Nanocrystal Memory Devices Self-Assembled by in Situ Rapid Thermal Annealing of Ultrathin a-Si on Sio2. Electrochem. Solid-State Lett. 2007, 10, H302-H304. (16) Borowik, L.; Barbe, J.-C.; Bussmann, E.; Cheynis, F.; Leroy, F.; Mariolle, D.; Muller, P. Method for Making Semi-Conductor Nanocrystals. U. S. Patent 8,647,957 B2, 2014. (17) Borowik, L.; Barbé, J.-C.; Bussmann, E.; Cheynis, F.; Leroy, F.; Mariolle, D.; Muller, P. Method for Making Semiconductor Nanocrystals Oriented Along a Predefined Direction. U. S. Patent 8,697,548 B2, 2014. 25 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 37

(18) Navas, H.; Maruyama, B.; Weaver, K.; Paillet, M.; Zahab, A.-A.; Fossard, F.; Loiseau, A.; Quesnel, E.; Anglaret, E.; Jourdain, V. Interplay of Interfacial Compounds, Catalyst Thickness and Carbon Precursor Supply in the Selectivity of Single-Walled Carbon Nanotube Growth. Carbon 2014, 80, 599-609. (19) Danielson, D. T.; Sparacin, D. K.; Michel, J.; Kimerling, L. C. Surface-Energy-Driven Dewetting Theory of Silicon-on-Insulator Agglomeration. J. Appl. Phys. 2006, 100, 083507. (20) Ishikawa, Y.; Imai, Y.; Ikeda, H.; Tabe, M. Pattern-Induced Alignment of Silicon Islands on Buried Oxide Layer of Silicon-on-Insulator Structure. Appl. Phys. Lett. 2003, 83, 3162-3164. (21) Leroy, F.; Cheynis, F.; Passanante, T.; Müller, P. Dynamics, Anisotropy, and Stability of Silicon-on-Insulator Dewetting Fronts. Phys. Rev. B 2012, 85, 195414. (22) Leroy, F.; Cheynis, F.; Passanante, T.; Müller, P. Influence of Facets on Solid State Dewetting Mechanisms: Comparison between Ge and Si on Sio 2. Phys. Rev. B 2013, 88, 035306. (23) Cheynis, F.; Leroy, F.; Müller, P. Dynamics and Instability of Solid-State Dewetting.

2013, 14, 578-589. (24) Nuryadi, R.; Ishikawa, Y.; Ono, Y.; Tabe, M. Thermal Agglomeration of SingleCrystalline Si Layer on Buried Sio2 in Ultrahigh Vacuum. J. Vac. Sci. Technol. B 2002, 20, 167-172. (25) Leroy, F.; Saito, Y.; Cheynis, F.; Bussmann, E.; Pierre-Louis, O.; Müller, P. Nonequilibrium Diffusion of Reactive Solid Islands. Phys. Rev. B 2014, 89, 235406.

26 ACS Paragon Plus Environment

Page 27 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(26) Leroy, F.; Saito, Y.; Curiotto, S.; Cheynis, F.; Pierre-Louis, O.; Müller, P. Shape Transition in Nano-Pits after Solid-Phase Etching of Sio2 by Si Islands. Appl. Phys. Lett.

2015, 106, 191601. (27) Sudoh, K.; Naito, M. Interfacial Reaction During Dewetting of Ultrathin Silicon on Insulator. C. R. Physique 2013, 14, 601-606. (28) Sudoh, K.; Naito, M. Interfacial Reaction of Si Islands on Sio2 During HighTemperature Annealing. J. Appl. Phys. 2010, 108, 083520. (29) Legrand, B.; Agache, V.; Nys, J. P.; Senez, V.; Stievenard, D. Formation of Silicon Islands on a Silicon on Insulator Substrate Upon Thermal Annealing. Appl. Phys. Lett. 2000, 76, 3271-3273. (30) Legrand, B.; Agache, V.; Mélin, T.; Nys, J. P.; Senez, V.; Stievenard, D. Thermally Assisted Formation of Silicon Islands on a Silicon-on-Insulator Substrate. J. Appl. Phys.

2002, 91, 106-111. (31) Borowik, L.; Chevalier, N.; Mariolle, D.; Martinez, E.; Bertin, F.; Chabli, A.; Barbé, J.C. Engineering the Size and Density of Silicon Agglomerates by Controlling the Initial Surface Carbonated Contamination. Appl. Surf. Sci. 2013, 270, 295-300. (32) Borowik, L.; Chevalier, N.; Mariolle, D.; Bertin, F.; Chabli, A.; Barbé, J.-C. The Influence of Strain on Dewetting of Silicon Films. J. Appl. Phys. 2013, 114, 063502. (33) Montoro, L. A.; Leite, M. S.; Biggemann, D.; Peternella, F. G.; Batenburg, K. J.; Medeiros-Ribeiro, G.; Ramirez, A. J. Revealing Quantitative 3d Chemical Arrangement on Ge-Si Nanostructures. J. Phys. Chem. C 2009, 113, 9018-9022.

27 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 37

(34) Lang, D. V.; People, R.; Bean, J. C.; Sergent, A. M. Measurement of the Band Gap of Gexsi1- X/Si Strained-Layer Heterostructures. Appl. Phys. Lett. 1985, 47, 1333-1335. (35) Schäffler, F. High-Mobility Si and Ge Structures. Semicond. Sci. Technol. 1997, 12, 1515. (36) Sant, S.; Lodha, S.; Ganguly, U.; Mahapatra, S.; Heinz, F. O.; Smith, L.; Moroz, V.; Ganguly, S. Band Gap Bowing and Band Offsets in Relaxed and Strained Si1- Xgex Alloys by Employing a New Nonlinear Interpolation Scheme. J. Appl. Phys. 2013, 113, 033708. (37) Pouch, S.; Amato, M.; Bertocchi, M.; Ossicini, S.; Chevalier, N.; Mélin, T.; Hartmann, J.-M.; Renault, O.; Delaye, V.; Mariolle, D., et al. Work Function Measurement of Silicon Germanium Heterostructures Combining Kelvin Force Microscopy and X-Ray Photoelectron Emission Microscopy. J. Phys. Chem. C 2015, 119, 26776-26782. (38) Hutin, L.; Cassé, M.; Le Royer, C.; Damlencourt, J.-F.; Pouydebasque, A.; Xu, C.; Tabone, C.; Hartmann, J.-M.; Carron, V.; Grampeix, H., et al. 20nm Gate Length Trigate Pfets on Strained Sgoi for High Performance Cmos. VLSI tech. Symp. 2010, 37-38. (39) Cheng, K.; Khakifirooz, A.; Loubet, N.; Luning, S.; Nagumo, T.; Vinet, M.; Liu, Q.; Reznicek, A.; Adam, T.; Naczas, S., et al. High Performance Extremely Thin Soi (Etsoi) Hybrid Cmos with Si Channel Nfet and Strained Sige Channel Pfet. Proceeding of Electron Devices Meeting 2012, 18.1.1-18.1.4. (40) Khakifirooz, A.; Cheng, K.; Nagumo, T.; Loubet, N.; Adam, T.; Reznicek, A.; Kuss, J.; Shahrjerdi, D.; Sreenivasan, R.; Ponoth, S., et al. Strain Engineered Extremely Thin Soi (Etsoi) for High-Performance Cmos. VLSI Tech. Symp. 2012, 117-118.

28 ACS Paragon Plus Environment

Page 29 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(41) Glowacki, F.; Le Royer, C.; Morand, Y.; Pédini, J.-M.; Denneulin, T.; Cooper, D.; Barnes, J.-P.; Nguyen, P.; Rouchon, D.; Hartmann, J.-M., et al. Ultrathin (5nm) Sige-onInsulator with High Compressive Strain (- 2gpa): From Fabrication (Ge Enrichment Process) to in-Depth Characterizations. Solid-State Electron. 2014, 97, 82-87. (42) Tezuka, T.; Sugiyama, N.; Takagi, S. Fabrication of Strained Si on an Ultrathin Sigeon-Insulator Virtual Substrate with a High-Ge Fraction. Appl. Phys. Lett. 2001, 79, 17981800. (43) Vincent, B.; Damlencourt, J. F.; Rivallin, P.; Nolot, E.; Licitra, C.; Morand, Y.; Clavelier, L. Fabrication of Sige-on-Insulator Substrates by a Condensation Technique: An Experimental and Modelling Study. Semicond. Sci. Technol. 2007, 22, 237. (44) Borowik, L.; Chevalier, N.; Mariolle, D.; Martinez, E.; Bertin, F.; Chabli, A.; Barbé, J.C. Controlled Synthesis of Sic Nanoparticles from Surface Silicon Contamination. 2013, 527, 133-136. (45) Bernard, J. E.; Zunger, A. Strain Energy and Stability of Si-Ge Compounds, Alloys, and Superlattices. Phys. Rev. B 1991, 44, 1663. (46) Garcia, R.; San Paulo, A. Attractive and Repulsive Tip-Sample Interaction Regimes in Tapping-Mode Atomic Force Microscopy. Phys. Rev. B 1999, 60, 4961. (47) Pavlyak, F. Modified Relative Sensitivity Factors for Auger Spectra. Appl. Surf. Sci.

1994, 81, 351-356. (48) Capellini, G.; Ciasca, G.; De Seta, M.; Notargiacomo, A.; Evangelisti, F.; Nardone, M. Agglomeration Process in Thin Silicon-, Strained Silicon-, and Silicon Germanium-onInsulator Substrates. J. Appl. Phys. 2009, 105, 093525. 29 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 37

(49) Berbezier, I.; Aouassa, M.; Ronda, A.; Favre, L.; Bollani, M.; Sordan, R.; Delobbe, A.; Sudraud, P. Ordered Arrays of Si and Ge Nanocrystals Via Dewetting of Pre-Patterned Thin Films. J. Appl. Phys. 2013, 113, 064908. (50) Zhang, P. P.; Yang, B.; Rugheimer, P. P.; Roberts, M. M.; Savage, D. E.; Liu, F.; Lagally, M. G. Influence of Germanium on Thermal Dewetting and Agglomeration of the Silicon Template Layer in Thin Silicon-on-Insulator. J. Phys. D Appl. Phys. 2009, 47, 175309. (51) Malachias, A.; Kycia, S.; Medeiros-Ribeiro, G.; Magalhaes-Paniago, R.; Kamins, T. I.; Williams, R. S. 3d Composition of Epitaxial Nanocrystals by Anomalous X-Ray Diffraction: Observation of a Si-Rich Core in Ge Domes on Si (100). Phys. Rev. Lett. 2003, 91, 176101. (52) Malachias, A.; Stoffel, M.; Schmidbauer, M.; Schulli, T. U.; Medeiros-Ribeiro, G.; Schmidt, O. G.; Magalhaes-Paniago, R.; Metzger, T. H. Atomic Ordering Dependence on Growth Method in Ge: Si (001) Islands: Influence of Surface Kinetic and Thermodynamic Interdiffusion Mechanisms. Phys. Rev. B 2010, 82, 035307. (53) Medeiros-Ribeiro, G.; Malachias, A.; Kycia, S.; Magalhaes-Paniago, R.; Kamins, T. I.; Williams, R. S. Elastic Energy Mapping of Epitaxial Nanocrystals. Appl. Phys. A 2005, 80, 1211-1214. (54) Ogawa, Y.; Toizumi, T.; Minami, F.; Baranov, A. V. Nanometer-Scale Mapping of the Strain and Ge Content of Ge/Si Quantum Dots Using Enhanced Raman Scattering by the Tip of an Atomic Force Microscope. Phys. Rev. B 2011, 83, 081302. (55) Hourani, W.; Periwal, P.; Bassani, F.; Baron, T.; Patriarche, G.; Martinez, E. Nanoscale Elemental Quantification in Heterostructured Sige Nanowires. Nanoscale 2015, 7, 8544-8553.

30 ACS Paragon Plus Environment

Page 31 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

(56) Shimizu, R.; Everhart, T. E.; MacDonald, N. C.; Hovland, C. T. Edge Effect in HighResolution Scanning Auger-Electron Microscopy. Appl. Phys. Lett. 1978, 33, 549-551.

31 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 37

TOC graphic

32 ACS Paragon Plus Environment

Page (a) 33 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13

The Journal of Physical (b) Chemistry

ACS Paragon Plus Environment

1 µm

500 nm

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 (c) 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Beginning state (b) ChemistryIntermediate The Journal of Physical Page 34 of 37

0.5 µm Fully dewetted

0.5 µm

(d) Number of dewetted agglomerates (a.u)

(a)

Diameter (nm) 100

ACS Paragon Plus Environment

1 µm

200

300 Diameter distribution Fit (2 gaussians)

HAADF The Germanium Chemistry (b)Journal of Physical

Page (a) 35 of 37

Silicon

(c)

Agglomerate

(d) 0

40 nm

40 nm

40 nm

Position from agglomerate surface (nm) 15 30 45 60 75 90 105

40 Germanium at. concentration (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

Box

Groove

~ 35%

30

BOx

SiGe Material

~ 22%

20

~ 21%

10 Raw film (Auger) Film's edge (Auger) Agglomerate (EDX)

0

ACS Paragon Plus Environment

0

2

4

6 8 10 12 Sputtering time (min)

14

16

(a)

. . B

D

C

.

.

Page SEM36 of 37

. 500 nm

500 nm Ge

(d)

Ge

500 nm

500 nm Si

(f)

Si

500 nm

500 nm

Mean diameter (nm)

Agglomerate Index B

C

D

E

220 200 180 160 140 120 100 80

60

2.2 2.0 1.8 1.6

5

(h) 3

ACS Paragon Plus Environment 1

NCs

NCs' edge

BOx

180

1.4

4

160

2

140

120

Mean diameter (nm)

100

1.2 1.0 Area 1 Area 2

80

0.8

IGe/ISi (a.u.)

6

IGe/ISi (a.u.)

A 1 2 3 4 5 6 E 7 8 9 (c) 10 11 12 13 14 15 16 17 18 19 20 (e) 21 22 23 24 25 26 27 28 29 30 31 32 A 33 341.0 35 360.8 37 380.6 39 400.4 41 42 430.2 (g) 44 450.0 46 Film

The Journal of Physical SEM (b) Chemistry

1 2 3 4 5 6 7 8 9 10 T° 11 12 13 14 15 16 17 18 19

500 nm

= 900°C

Ge concentration in nanocrystals

Decreasing The Journal of Physical Chemistry

Page of 37 T° = 37 750°C

SiO2

SiGe-NC

SiO2

ACS Paragon Plus Environment

500 nm

diameters

SiO2

SiO2

Interfacial reaction advencement