Growth of In-Plane Ge1âxSnx Nanowires with 22 at. % Sn Using a

Subscriber access provided by University of Sunderland

C: Physical Processes in Nanomaterials and Nanostructures 1-x

x

Growth of In-Plane Ge Sn Nanowires with 22 at. % Sn Using Solid-Liquid-Solid Mechanism Edy Azrak, Wanghua Chen, Simona Moldovan, Shiwen Gao, Sébastien Duguay, Philippe Pareige, and Pere Roca i Cabarrocas J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.8b07142 • Publication Date (Web): 23 Oct 2018 Downloaded from http://pubs.acs.org on October 24, 2018

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 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 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.

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 23 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

Growth of In-Plane Ge1-xSnx Nanowires With 22 at.% Sn Using Solid-Liquid-Solid Mechanism Edy Azrak1, Wanghua Chen2,3 *, Simona Moldovan1, Shiwen Gao2, Sébastien Duguay1, Philippe Pareige1* and Pere Roca i Cabarrocas2 1GPM,

Université et INSA de Rouen, CNRS, Normandie Université, 76801 Saint Etienne du Rouvray, France

2LPICM, 3Faculty

CNRS, Ecole polytechnique, Université Paris-Saclay, 91128 Palaiseau, France of Science, Ningbo University, 315211 Ningbo, China

*E-mail: [email protected]; [email protected] Abstract Germanium-tin alloys have gained strong attention because of their optical and electrical properties and their compatibility with silicon-based technologies. By increasing the Sn content in the alloy, the charge carrier mobility can be improved and the energy bandgap can be transformed from indirect to direct. However, the fabrication of GeSn is a huge challenge as the equilibrium solubility of Sn in Ge is limited to < 1 at. %. The aim of this study is the fabrication of in-plane GeSn nanowires catalyzed by Sn nanoparticles to overcome this limit. TEM-based analysis show that the in-plane GeSn nanowires can reach an out of equilibrium Sn concentration of 22 at. %. Moreover, a homogenous incorporation of Sn into the Ge nanowires is achieved during the in-plane solid-liquid-solid growth process, without Sn segregation or precipitation in the host material. The results are discussed in the framework of three kinetic1 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

based models: the step growth model, the continuous growth model, and the dimer insertion model, to account for the non-classical physical aspects behind the extraordinary catalyst incorporation.

Introduction Over the past decades the performance of Si ultra-large-scale integrated circuits has improved by increasing the number of elementary components (i.e. transistors, diodes). Further improvements in Si-based electronics and optoelectronics are limited by the crystalline Si properties, such as the carrier mobility and the indirect band-gap 1. As a viable alternative to Si, the group IV compound Germanium-Tin (Ge1-xSnx) is a unique class semiconductor offering adjustable electrical and optical properties as a function of Sn concentration. This binary alloy presents many interesting properties such as: 1) high hole mobility2 compared to Si 3 and Ge 4 individually, and can be improved by increasing the Sn content, 2) tunable bandgap 5, 3) direct bandgap transition for a sufficiently high Sn content, 4) low thermal budget required to grow a GeSn alloy, due to its eutectic point (231.1℃), and 5) strain engineering as a stressor for Ge 6. The Ge1-xSnx is very suitable as a channel material for high speed and low power consumption transistors 7. In fact, the incorporation of Sn modifies the electronic band structure of the system. For instance in Ge1-xSnx alloys, the direct Γ-valley decreases faster than the indirect valley (L) in the conduction band (at k = 0) with the increase of Sn incorporation. The energy difference between Γ and L valleys in c-Ge is ∆𝐸 = 0.14 eV. When the Γ-valley crosses-over the L-valley the material becomes a direct gap 8. By using the nonlocal empirical pseudopotential method 9, it has been reported that the required Sn concentration to transform Ge into a direct gap semiconductor is above 6 %, although some 2 ACS Paragon Plus Environment

Page 2 of 23

Page 3 of 23 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


discrepancies regarding this value exist

10-12.

However, to fabricate GeSn nanowires (NWs)

having such concentration, several fundamental obstacles have to be solved: i) the solubility of Sn in Ge is only 1% below 500℃

13

, ii) the strong Sn segregation behavior

14

and iii) the

larger size of Sn atoms compared to Ge 15, which can degrade the Ge lattice. Extensive efforts have been done to break the barrier of the Sn equilibrium concentration without degrading the crystalline quality of GeSn alloys. Recently, a successful GeSn thin film has been grown with relatively high Sn content (~ 17% at. Sn) using molecular beam epitaxy (MBE) at low temperature (~ 140°C), but failed in suppressing the phase separation tendency16. The nanowire configuration has also been investigated because of its high surface-to-volume ratio allowing a better strain-relaxation compared to the thin-film approach17. Among the most popular growth techniques18-22 (vapor-liquid-solid, vapor-solidsolid, and electrodeposition), the most effective mechanism for impurity incorporation is the one that allows the growth under non-equilibrium conditions23. For instance, Au-catalyzed Ge NWs grown by vapor-liquid-solid (VLS) mechanism could reach ~ 10% at. Sn24. Interestingly, it has been shown by Chen et al.25 that the catalyst incorporation into nanowires strongly depends on their growth rate. By comparing the vapor liquid solid (VLS) (< 10 nm.s1)

and in-plane solid liquid solid (IPSLS) (~ 40 nm.s-1) growth rates for In and Sn-catalyzed Si

NWs, the SLS exhibits a higher growth rate. Hence, much higher catalyst incorporation can be achieved for the IPSLS process. Another reason to use the IPSLS as a growth mode for GeSn NWs, from an application point of view, is the difficulty of integrating vertical (out-ofplane) NWs grown by VLS into 2D planar devices. A formidable challenge is to manipulate and assemble vertical NWs into a 2D layout for large-scale integration of NW-based devices. The NWs fabricated by the IPSLS are readily grown in-plane, no additional manipulation is

3 ACS Paragon Plus Environment


Page 4 of 23

required for implementing them in planar devices. Furthermore, in-plane NWs can be guided during their growth, which provides a general strategy to deploy the NWs precisely into a desired circuit 26. In this study, we focus on overcoming the fabrication difficulties of in-plane crystalline GeSn NWs with a high and homogeneously distributed Sn content. The study also covers the thermal stability of the GeSn NWs. Two Sn-based catalysts are used for the growth: 1- metallic Sn nanoparticles (NPs), and 2- SnO2 colloids. The in-plane solid-liquid-solid (IPSLS) method is used27 for the growth, by employing a thin film of hydrogenated amorphous Ge (a-Ge:H) as the source of Ge atoms for GexSn1-x NWs growth. The samples are grown at 270°C and 350°C. Detailed characterizations based on micro-Raman spectroscopy and scanning transmission electron microscopy (STEM) coupled with energy dispersive spectroscopy (EDS) are performed. The high incorporation of Sn into the Ge NWs is discussed in the framework of three kinetic models: step growth model (SGM)

28,

continuous growth model (CGM)

29,

and

dimer-insertion model (DIM) 30.

Experimental description The first set of growth experiments is carried out using pure Sn NPs as catalysts, with an average diameter of 16 nm (Fig.S1a,d in SI – 1). These Sn NPs are deposited thermally on a Si substrate. When the sample is introduced in the PECVD (plasma-enhanced chemical vapor deposition), a hydrogen plasma is used to chemically reduce the oxide layer formed on Sn NPs. Then, an a-Ge:H layer is deposited on top of the Sn NPs, and the temperature is raised above the melting point of Sn (~ 232°C) to activate the GeSn NW growth without using any plasma during this stage (see SI – 1 for more details on the growth conditions). Growth attempts using


Page 5 of 23 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


this strategy were unsuccessful, possibly due to the high density of Sn NPs (1.9×1011/cm2 as shown in Fig.S1). Sn NPs should be distant from each other, since the growth of NWs requires a space where a NP can travel to do the growth. In order to make room for the NPs to do their trip, Sn strips are used (see Fig.S2). This sample has a space between the Sn strips where NPs can travel while growing GeSn NWs. However, even though between the Sn strips there is enough space with only a-Ge:H on top of the Si substrate, the growth was unsuccessful (as shown in Fig.S2). After careful analysis, it can be concluded that the strong wetting of Sn on a-Ge:H surface is the main blocking factor (Fig.S2.b). Indeed, Sn is known to be an unstable catalyst

31

due to its low surface tension 32,

especially when it is in contact with a-Ge:H 33, a high surface energy material. In addition, Sn NPs have a higher wettability on a-Ge:H and a higher solubility in the corresponding layer 34 as compared to a-Si:H35. For this reason, the growth of in-plane GeSn NWs using Sn is certainly more challenging than the growth of Si NWs. To overcome this drawback, SnO2 NPs were used as a catalyst precursor. The aim was to limit the wetting behavior of pure Sn and prevent its splitting into smaller NPs. The SnO2 NPs are partially reduced to pure Sn using a H2 plasma, as shown in Fig.1a. Afterwards, the same condition of a-Ge:H deposition as in the previous trials is used (Fig.1b). During the growth at 270°C, the SnO2 remains in its solid state due to its high melting point (> 1600°C). At the initial stage of growth, the Ge concentration in Sn NP starts to increase and the SnO2 NP plays the role of a ledge that prevents the liquid Sn from splitting and wetting until supersaturation of Ge in the Sn NP is attained. Energetically favorable sites for crystalline nanowire nucleation are located at the Sn – SnO2 interface. Many crystal seeds can form, but the one having the lowest surface-tovolume ratio, thus, the lowest Gibbs energy, will preferentially attract the incoming Ge atoms. 5 ACS Paragon Plus Environment


Once a catalyst-seed interface is built, an absorption and deposition edge will appear. Amorphous Ge has a higher Gibbs energy than its crystalline form; this energy difference, between the back side (catalyst-crystalline interface) and the front side (catalyst-amorphous interface) of Sn NPs, creates a thermodynamic driving force that drags the catalysts towards the amorphous layer as shown in Fig.1c. Since Sn NPs are liquid, they dissolve the a-Ge:H at their front absorption interface creating a trench during their movement on the substrate. Note that Sn NPs advance towards a-Ge:H by slightly wetting the a-Ge:H thin film, but they do not completely spread (full-wetting) otherwise the NW growth would fail. The trench is an extremely thinned (or totally consumed) a-Ge:H region, which provides a trace of a Sn NP diameter on the a-Ge:H layer. As a consequence, the total system energy is reduced upon the transformation of a-Ge:H into a c-Ge NW. As can be inferred by Fig.1d, the growth of GeSn NWs is achieved successfully, and the liquid Sn NPs did not wet when they got detached from the SnO2 NPs. In fact, the success trials (using SnO2 precursor) have shown that we succeeded in holding the Sn NPs from wetting and to reach the supersaturation. The SnO2 acts as a solid edge that prevents the Sn NP from wetting. Therefore, the liquid Sn needs to be in contact with a solid edge in order not to wet. For this reason, we believe that the existence of the nanowire behind the catalyst is what preventing the catalyst from complete wetting during the growth. An IPSLS GeSn is shown in Fig.1d. A general view (low magnification) of GeSn NWs is shown in the supplementary information (SI – 2).


Page 6 of 23

Page 7 of 23 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


Figure 1 Growth studies using SnO2 NPs. a) Schematic illustration showing the reduction step of SnO2 NP partially into pure Sn using a H2 plasma at 250℃. b) A 14 nm of a-Ge:H layer is deposited on top of the system at 120℃. c) Growth of GeSn NW at 270℃, the thermodynamic force, 𝐹, drags the liquid Sn NP to dissolve the a-Ge:H and achieve the growth. d) SEM top view of a GeSn NW grown by the SLS mechanism. The SnO2 NP remains at the initial position. The trench is an empty region or an extremely thinned a-Ge:H, it is created by the catalyst. The end of the growth is marked by a wetting behavior of the catalyst.

Characterizations Raman spectroscopy. As a first approach to characterize the composition of GeSn NWs, micro-Raman spectroscopy was performed. Fig. 2 shows the micro-Raman spectra measured on an a-Ge:H layer and on a region containing GeSn NWs.

Bulk Ge has a strong Ge-Ge

longitudinal optical (LO) phonon peak at 𝑤𝐺𝑒 = 301 ± 0.5 𝑐𝑚 ―1 36, which progressively shifts to lower frequencies with an increasing Sn concentration. The wavenumber shift can be 7 ACS Paragon Plus Environment


described by a linear equation: ∆𝑤𝐺𝑒𝑆𝑛 = ― 𝑘.𝑥 𝑐𝑚 ―1 where x is the Sn atomic fraction and k is a prefactor depending on the strain state of the nanowires37. Accordingly, for strained and relaxed nanowires, the prefactors are respectively 𝑘𝑠 = 30.3 and 𝑘𝑟 = 76.8. In Figure 2 the GeGe longitudinal optical (LO) phonon peak is located at 293 𝑐𝑚 ―1 due to the spectral feature of the compressive strain and the substitution-Sn-induced bond stretching in GeSn NWs. Using the linear shift equation, the Sn atomic percentage in GeSn NWs is estimated to be ~ 10 % for a completely relaxed structure and ~ 26% for a strained structure.

Figure 2 Incorporation of Sn in GeSn nanowires. Micro-Raman spectra (λ = 633 nm excitation) of a sample containing GeSn NWs (blue) and a region containing only a-Ge:H (red).

TEM charaterizations. Transmission Electron Microscopy (TEM) is used to give valuable insights on GeSn NWs structure and chemistry. In Scanning TEM (STEM) mode, and by using High Angular Annular Dark Field (HAADF), the intensity scales with the atomic number of the constitutive elements. This allows complementary analysis with Energy 8 ACS Paragon Plus Environment

Page 8 of 23

Page 9 of 23 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


Dispersive Xray spectroscopy (EDX) to map the chemical composition. The complete STEMHAADF and STEM-EDS investigation of a GeSn NW cross-section prepared by Focused Ion Beam (FIB) is displayed in Figure 3. According to the linear concentration profile of the 55 nm diameter NW shown in Fig.3c, the Ge and Sn have a spatial distribution complementarity. The Ge varies from 20 at.% which marks the edge of the NW to reach 70 at.% inside. Whereas, in Fig.3d, the EDX cartography shows that Ge and Sn are homogeneously distributed at the nanometer scale without visible precipitations. Since a 1.5 nm spatial resolution has been employed for the STEM-EDS mapping experiments, the presence of atomically ordered GeSn alloy cannot be ascertained. For a more accurate quantification, a well-defined rectangular area was selected for an EDX count, the average Sn composition is 22 ± 3 at.% Sn. These measurements have been repeated for other GeSn NWs produced under the same growth conditions, therefore leading to an average concentration of 21.3 at.% Sn (see supplementary information SI – 3). The high resolution HAADF micrograph (in Fig.3b) identifies the crystalline nature of the NW. As shown by the corresponding FFT (inset in the high resolution image), the NW is oriented along the [220] direction. This orientation has been also confirmed by the direct measurement of the distance between atoms in the image (see SI – 4). By correlating the STEM-EDS analysis and the micro-Raman of the Sn concentration, we can infer that GeSn NW is strained (the prefactor is k = 36.4) due to the accommodation of ~ 22 % of Sn atoms in the Ge lattice.



Figure 3 High-resolution two-dimensional Energy Dispersive X-ray (EDX) map and structure of a Sn catalyzed Ge nanowire. a) HAADF micrograph showing a radial cross-section of a GeSn NW grown on a Si substrate at 270°C, and protected by a Pt layer. b) A high-resolution image, depicted as HR Zone (white square), shows the crystalline nature of the material. c) The composition profile along the black arrow (along the diameter of the nanowire) indicates the atomic percentage of Ge and Sn (the complement to 100% is due to presence of Si, Ga, O, and Pt atoms incorporated at the surface of the sample during the ion beam preparation process). d) A cross-sectional HRTEM – EDX, 30 min count, displays the elemental spatial distributions of Ge in red, Sn in green, Si in blue, and Pt in yellow.


Page 10 of 23

Page 11 of 23 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


Discussion The obtained concentration of Sn in GeNWs is highly out-of-equilibrium; it is roughly 44 times more for than the solubility limit (0.5 at.% Sn) in Ge at the growth temperature used (270℃). Different models28-30,38 have been developed to explain the shift in the equilibrium solubility of a solute in a host material. The difference between them is the atomic motion of the growing interface between the solid (i.e. NW) and the liquid (i.e. Sn NP). 1. Step Growth Model (SGM). In order to explain the high incorporation of Sn in GeSn NWs, one imperatively needs to address the problem of the atomic motion within the system at the NW/catalyst interface. The main model used to describe the nanowires growth is based on a step growth process23,24. In this regime, the growth is attributed to the progressive addition of atomic monolayers at the solid side of the NW/catalyst interface. The interface advances by the lateral flow of atomic steps. The rapid progression of the NW/catalyst interface traps the solute atoms (Sn in our case) escaping from the solid side of this interface (i.e. GeSn NW), which diffuse slower than the moving interface; this phenomenon is known as the “solute trapping”. The concentration of solute atoms undergoes a strong deviation towards higher values from the classical solubility limit upon a growth rate of few m/s39. Amazingly, in our case, for 22 at.% Sn incorporation, the growth rate is of few tens of nm/s, which is much lower than the capability of kinetic trapping as predicted by the existing model39. Biswas et al.24 suggested that the time required for Sn atoms to escape from the solid interface towards the liquid corresponds to the time needed to complete a step. However, at the atomic level, the process consists of embedding Sn atoms at the step-edges (at the solid part of the NW/catalyst interface) right after the construction of the next row of atoms. Therefore, the velocity of atomic row formation Val (see Fig. SI.6) should be accounted for this non-classical incorporation of Sn rather than the step 11 ACS Paragon Plus Environment


Page 12 of 23

velocity Vstep. Despite these considerations, the use of step growth model predicts an insignificant Sn incorporation at the interface (see SI – 5). 2. Continuous growth model (CGM). This growth mechanism (CGM) is another common kinetic model that might account for the incorporation of Sn beyond the equilibrium limit 40. This model operates in the case of a high driving force (conversely to step-flow kinetics) where the solid part of the interface (NW/catalyst) will advance continuously 41. The CGM model treats the case when the interface is atomically rough. In a steady-state, the growth occurs by direct impingement of atoms on high energy sites of the interface

42.

However, a detailed atomistic

description of the solute incorporation process is not provided. By inserting Vint (that is the growth rate) into the CGM equation (shown in SI – 6), the velocity dependent partition coefficient gives an insignificant value. In fact, the solute atoms (Sn) have a diffusive velocity vD ~ 5 m/s which is fast enough compared to Vint to be trapped at the interface. The Val cannot be used in this model since it is a direct impingement mechanism. Even if the Val is used in the equation, the partition coefficient is ~ 10-5. Therefore, these kinetic models (step and continuous growth) only fit the rapid cooling process where growth rates are few m/s, which are necessary to trap the diffusive solute atoms. 3. Dimer-insertion model (DIM). For the case of Sn incorporation into Si NWs, Chen et al. proposed a model based on the insertion of dimers25, where the nucleation of Si is achieved via Si pairs or dimers (Si-Si) and the incorporation of Sn atoms is realized via Si-Sn dimers. At low growth temperature (e.g. < 400 °C), the dimers are likely to be stable without breaking into two atoms. By analogy, the same dimer-based nucleation mechanism can be proposed for our GeSn NWs. Once the Ge-based dimers are formed, other adatoms or dimers can join and constitute the nuclei of a growing terrace. Such kind of nucleation around the Ge-based dimers 12 ACS Paragon Plus Environment

Page 13 of 23 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


occurs on the whole growing interface with multi sites. As a consequence, growing terraces can coalesce with each other and then form a complete layer. Based on the dimer nucleation with multi sites, a very high interface velocity (NW growth rate) is expected. Two types of dimers can be found in our NW growth system including Ge-Ge and Ge-Sn. With the increase of NW growth rate (higher supersaturation), more Ge adatoms in the Ge/Sn alloy droplet are available allowing more Ge-Sn dimers to be formed, therefore higher Sn incorporation. Chen et al.25 showed that high growth rates lead to high Sn and In incorporation into SiNWs grown by the IPSLS method. This is due to the direct contact between Sn NPs with the a-Ge:H layer, which constitutes a large source of Ge atoms and creates a powerful force dragging the catalyst, and thus leads to higher growth rates compared to VLS. To speculate the Sn concentration incorporated in the GeSn NWs, it is crucial to determine the growth rate. As a matter of fact, while raising the temperature to reach the desired value, the NWs start to grow as soon as the supersaturation is reached in the liquid catalyst. When reaching 270℃, many GeSn NWs have already grown at lower temperatures (e.g. 235℃). This means that the exact temperature at which a NW is grown is difficult to determine. Each time the temperature increases by 1℃ to reach the desired value (i.e. 270℃), group of catalysts having a certain size are activated. As a consequence, the GR is hard to determine since the temperature at which the NWs start to grow is not known. Therefore, dividing the length of a NW by the annealing duration (which should be the growth duration) will probably give a false GR value. Thus, the best approach is to consider an interval of GRs. To determine a feasible range within our framework, we consider the average length of our IPSLS GeSn NWs (1 μm) and that the growth occurs between 1 min and 3 min. Therefore, the estimated GR interval is [5 nm/s – 16 nm/s]. The concentration of Sn, CSn, is given by: 13 ACS Paragon Plus Environment


𝐶𝑆𝑛 =

𝜈𝐺𝑒 ― 𝑆𝑛 𝜈𝐺𝑒 ― 𝐺𝑒 + 𝜈𝐺𝑒 ― 𝑆𝑛

(2)

Where 𝜈𝐺𝑒 ― 𝐺𝑒 and 𝜈𝐺𝑒 ― 𝑆𝑛 are the incorporation rates of Ge-Ge and Ge-Sn dimers respectively which are obtained by analogy to 𝜈𝑆𝑖 ― 𝑆𝑖 and 𝜈𝑆𝑖 ― 𝑆𝑛 25. The detailed analysis is provided in SI – 7. The Sn distribution is homogeneous throughout the NW with no visible segregation towards structural defects. From the model we can predict the incorporation of Sn in the GeSn NWs as shown in Fig. 4, where the black line represents the Sn at.% function of the growth rate. The temperature necessary to activate the Sn segregation is between 320℃ and 350℃, therefore the resultant uncertainty is 2 meV (see SI – 7 for details). As a consequence, the curve has a certain error margin (represented by two red curves) as shown in Fig.4. The error at low growth rates is smaller (∆ = 0.5 at.%) than at high growth rates (∆ = 1.8 at.%).. The concentration predicted is 22 ± 0.5 at.% Sn, and the corresponding velocity is 7 nm/s which relies in the previously determined interval of GRs [5 nm/s and 16 nm/s] for the IPSLS GeSn NWs grown at 270℃.


Page 14 of 23

Page 15 of 23 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


Figure 4 Dependence of Sn concentration on GeSn NWs growth rate (black curve) based on the dimer insertion model. The Sn atomic percentage values (red region) correspond to the GR interval [5 nm/s – 16 nm/s] (blue region) estimated for the growth of IPSLS GeSn NWs at 270℃. The yellow point on the curve shows that for a 22 at.% Sn incorporation the GR should be equal to 7 nm/s. The error caused by the uncertainty of the temperature at which the segregation starts (between 320℃ and 350℃) is represented by the two red curves.

Thermal stability. The technology steps to process GeSn NWs into electronic or optoelectronic devices will involve annealing steps (for instance for contact optimization). It is therefore important to determine a thermal window in which GeSn NWs are stable. Under thermal constraints, Ge-Sn systems are known to phase-separate. Annealing such a system might induce changes on its microstructural and physical properties. In the case of an epitaxially grown GeSn thin-film, strain relaxation starts by the formation of misfit dislocations 15 ACS Paragon Plus Environment


at temperatures between 450℃ - 540℃, whilst higher temperatures lead to a steeper segregation and precipitation43.

Figure 5. Thermal stability of GeSn NWs. a) SEM image showing a Ge0.78Sn0.22 NW grown at 270℃ without further annealing. b) SEM image showing NWs of the same sample shown in (a) but annealed at 350℃ during 3 hours. c) STEM – bright field micrograph showing a Z-contrast, where Sn atoms (heavier than Ge) are darker than Ge atoms. Note that this NW is protected by a Pt layer during FIB preparation. d) The STEM-EDX map displays the spatial distribution of Ge in red, Sn in green, and Si in blue, explaining the inhomogeneous radial contrast. Thus the EDX presents a clear evidence of a phase-separation into two phases: a Ge-rich region (< 1 at.% Sn) and a Sn-rich region (with at.% Ge smaller than the detectable threshold).

Upon annealing at 350℃, the NWs appear rougher compared to the unannealed ones, as shown be the SEM images in Fig.5a,b. For deep structural and composition insights, TEM 16 ACS Paragon Plus Environment

Page 16 of 23

Page 17 of 23 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


analysis are used. The phase separation starts at 350℃, which is lower than the separation temperature previously reported by Li et al. (for ~ 8 at.% Sn ). As shown in Fig.5c, two distinct regions (Ge and Sn) are obtained inside the nanowire. According to EDX measurements (Fig.5d) we identify a nearly pure Sn region (the Ge content here is below the detectable amount of 0.1 at.%), whereas Ge-rich regions contain a traceable amount of Sn (< 1 at.%). The bottom Ge region in Fig.5d does not appear well in the EDX map due to the superposition of Ge and Sn maps. Therefore, in order for this alloy to be stable, the temperature should be less than 350℃ during device fabrication. Conclusion and perspectives In summary, we have shown that the incomplete reduction of SnO2 to pure Sn is crucial for the successful growth of GeSn NWs. At the first stages of the growth, Sn NPs tend to wet and diffuse in the a-Ge:H layer before supersaturation occurs. At this stage, the contact between the Sn drop and the SnO2 NP reduces the wetting behavior of liquid Sn, which allows the supersaturation to be reached. Micro-Raman characterizations show a shift in Ge-Ge (LO) peak of ∆w = 8 cm-1 due to the incorporation of Sn into the Ge NW. From the shift we estimate a Sn concentration of ~ 10% or ~ 26% depending on whether the NW is fully relaxed or strained. TEM-based EDX measurements show an extraordinary Sn content (up to 22%) in GeSn NWs. The combination of EDX and the micro-Raman results indicates that the NWs are strained; and the prefactor is equal to 36.4. The spatial distribution of Ge and Sn within the NWs is homogeneous and no clusters were detected within the NWs cross-sections. Increasing the growth temperature leads to a Ge-Sn phase separation, such that the thermal window for a stable GeSn processing into devices is < 350℃. The step growth (SGM), and continuous (CG) growth models cannot describe the high concentration of Sn in our GeSn NWs produced at 17 ACS Paragon Plus Environment


growth rates of tens of nm/s. We propose that the growth mechanism of GeSn NWs follows the dimer-insertion model, which predicts reasonable values of Sn incorporated fraction corresponding to our GR interval. Methods Sn particles For evaporated Sn catalysts, high purity Sn powder is evaporated and deposited onto the surface of a Si substrate. Afterwards, the sample is introduced in a PECVD reactor. A radiofrequency H2 plasma is ignited at a substrate temperature of 400℃ in order to remove the native oxide from the Sn film and obtain Sn droplets. Then the temperature is decreased to 120℃ and a hydrogenated amorphous Germanium (a-Ge:H) layer ~ 14 nm is deposited on top of the catalyst NPs. The deposition of a-Ge:H is done under a flow rate of GeH4 and H2 of 50 and 100 SCCM (standard cubic centimeter per minute under STP) respectively. As a final step, the temperature is raised above the eutectic point of Sn (231℃, for instance 270℃) to activate the NWs growth. SnO2 colloids For SnO2 catalysis, 8 grams of commercial SnO2 powder are added to 25 ml of Diethylene glycol (DEG). The mixture is grinded in a milling machine to reduce the size of SnO2 particles and disperse them. By centrifugation lighter SnO2 are separated from the heavier ones. Tetramethylamonium Hydroxide (TMAOH) is used as a solvent to prevent precipitation, and an ultra-sonic treatment is applied to better disperse the NPs in the colloidal solution. The colloidal solution is deposited on a Si substrate and introduced in the PECVD reactor. The oxide catalyst is partially reduced to pure Sn using a H2 plasma for 10 min and a RF power density of 22


Page 18 of 23

Page 19 of 23 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


mW/cm2. For the a-Ge:H deposition and annealing procedure we used the same parameters as in the case of Sn evaporation.

Acknowledgements The authors would thank Isabelle Maurin, PMC, Ecole Polytechnique for preparing the SnO2 colloids, and Zhaoguo Xue, school of electronics science and engineering – Nanjing University, for preparing Sn strips. Supporting information The supporting Information is available free of charge on the ACS Publication website. Growth trials using evaporated Sn and Sn strips, a general view (at low magnification) of inplane SLS GeSn NWs, TEM micrographs of axial and radial cross-sections of in-plane SLS GeSn NWs, and their corresponding EDX cartographies and quantifications, crystal orientation identification of the NW using the atomic distance as measured by the high-resolution micrograph, step growth model calculations using the atomic line velocity, the partition coefficient as given by the continuous growth model, the dimer insertion model adjustment for the Sn-catalyzed GeSn NWs which gives the Sn concentration dependence on the growth rate or interface velocity.

References 1

Jacobsen, R. S.; Andersen, K.N.; Borel, B.I.; Pedersen, J.F.; Frandsen, L.H.; Hansen, O.; Kristensen, M.; Lavrinenko, A.V.; Moulin, G.; Ou, H.; et al. A. Strained Silicon as a New Electro-Optic Material. Nature, 2006, 441, 199-202. 19 ACS Paragon Plus Environment


2 3 4 5 6 7

8 9 10 11 12 13 14 15 16 17 18 19 20 21

Noriyuki, U.; Tatsuro, M.; Ruben, R. L; Shingo, O.; Yuji, O.; Ryohei, T.; Manabu, I.; Locquet, J.P. Carrier and Heat Transport Properties of Polycrystalline GeSn Films on SiO2. Appl. Phys. Lett., 2015, 107, 232105. Sau, J. D.; Cohen, M.L. Possibility of Increased Mobility in Ge-Sn Alloy System. Phys. Rev. B, 2007, 75, 045208. Ruben, R. L.; Tatsuro, M.; Wipakorn, J.; Hiroyuki, H.; Noriyuki, U.; Shu, M.; Masatoshi, T.; Locquet, J.P. Tensile-Strained GeSn Metal–Oxide–Semiconductor Field-Effect Transistor Devices on Si(111) Using Solid Phase Epitaxy. Appl. Phys. Express, 2013, 6, 101301. Cong, H.; Yang, F.; Xue, C.; Yu, K.; Zhou, L.; Wang, N.; Cheng, B.; Wang, Q. Multilayer Graphene–GeSn Quantum Well Heterostructure SWIR Light Source. Small, 2018, 14, 1704414. Zaima, S.; Nakatsuka, O.; Taoka, N.; Kurosawa, M.; Takeuchi, W.; Sakashita, M. Growth and Applications of GeSn-Related Group-IV Semiconductor Materials. Sci. Technol. Adv. Mater., 2015, 16, 043502. Wang, L.; Shaojian, S.; Wang, W.; Gong, X.; Yang, Y.; Guo, P.; Zhang, G.; Xue, C.; Cheng, B.; et al. Strained Germanium–Tin (GeSn) p-Channel Metal-Oxide-Semiconductor Field-EffectTransistors (p-MOSFETs) with Ammonium Sulfide Passivation. Solid-State Electron, 2013, 83, 66-70. Zhang, P.; Crespi, V. H.; Chang, E.; Louie, S.G.; Cohen, M.L. Computational Design of DirectBandgap Semiconductors that Lattice-Match Silicon. Nature, 2001, 409, 69-71. Gupta, S.; Magyari-Köpe, B.; Nishi, Y.; Saraswat, K. C. Achieving Direct Band Gap in Germanium Through Integration of Sn Alloying and External Strain. J. Appl. Phys., 2013, 113, 073707. Moontragoon, P.; Soref, R.A.; Ikonic, Z. The Direct and Indirect Bandgaps of Unstrained SixGe1−x−ySny and Their Photonic Device Applications. J. Appl. Phys., 2012, 112, 073106. Pairot, M.; Ikonić, Z.; Harrison, P. Band Structure Calculations of Si–Ge–Sn Alloys: Achieving Direct Band Gap Materials. Semicond. Sci. Technol., 2007, 22, 742. Gupta, S., Blanka, M. K.; Yoshio, N.; Krishna, C. S.; Achieving Direct Band Gap in Germanium Through Integration of Sn Alloying and External Strain. J. Appl. Phys., 2013, 113, 073707. Trumbore, F.A.; Solid Solubilities and Electrical Properties of Tin in Germanium Single Crystals. J. Electrochem. Soc., 1956, 103, 3. Li, H.; Chang, C.; Chen, T.P.; Cheng, H.H.; Shi, Z.W.; Chen, H. Characteristics of Sn Segregation in Ge/GeSn Heterostructures. Appl. Phys. Lett., 2014, 105, 151906. Gencarelli, F.; Vincent, B.; Demeulemeester, J.; Vantomme, A.; Moussa, A.; Franquet, A.; Kumar, A.; Bender, H.; Meersschaut, J.; Vandervorst, W.; et al. Crystalline Properties and Strain Relaxation Mechanism of CVD Grown GeSn. ECS Trans., 2013, 50, 875-883. Kormoš, L.; Kratzer, M.; Kostecki, K.; Oehme, M.; Šikola, T.; Kasper, E.; Schulze, J.; Teichert, C. Surface Analysis of Epitaxially Grown GeSn Alloys with Sn Contents Between 15% and 18%. Surf. Interface Anal., 2017, 49, 297-302. Xiang, H.J.; Wei, S.H.; Da Silva, J.L.F.; Li, J. Strain Relaxation and Band-Gap Tunability in Ternary InxGa1-xN Nanowires. Phys. Rev. B, 2008, 78, 193301. Karam, C.; Guerra-Nuñez, C.; Habchi R.; Herro Z.; Abboud N.; Khoury A.; Tingry S.; Miele P.; Utke I.; Bechelany M. Urchin-Inspired ZnO-TiO2 Core-Shell as Building Blocks for Dye Sensitized Solar Cells. Mater. Des., 2017, 126, 314-321. Greytak, A. B.; Lauhon, L. J.; Gudiksen, M. S.; Lieber, C. M. Growth and Transport Properties of Complementary Germanium Nanowire Field-Effect Transistors. Appl. Phys. Lett., 2004, 84, 4176-4178. Kodambaka, S.; Tersoff, J.; Reuter, M. C.; Ross, F. M.; Germanium Nanowire Growth Below the Eutectic Temperature. Science, 2007, 316, 729-732. Karam, C.; Habchi, R.; Tingry, S.; Miele, P.; Bechelany, M. Design of Multilayers of Urchin-like ZnO Nanowires Coated with TiO2 Nanostructures for Dye-Sensitized Solar Cells. ACS Appl. Nano Mater., 2018, 1, 3705–3714. 20 ACS Paragon Plus Environment

Page 20 of 23

Page 21 of 23 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


22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

37

38 39 40 41 42

Makhlouf, H.; Karam, C.; Lamouchi, A.; Tingry, S.; Miele, P.; Habchi, R.; Chtourou, R.; Bechelany, M.; Analysis of Ultraviolet Photo-Response of ZnO Nanostructures Prepared by Electrodeposition and Atomic Layer Deposition. Appl. Surf. Sci., 2018, 444, 253-259. Moutanabbir, O.; Isheim, D.; Blumtritt, H.; Senz, S.; Pippel, E; Seidman, D.N. Colossal Injection of Catalyst Atoms into Silicon Nanowires. Nature, 2013, 496, 78-82. Biswas, S.; Doherty, J.; Saladukha, D.; Ramasse, Q.; Majumdar, D.; Upmanyu, M.; Singha, A.; Ochalski, T.; Morris, M.A.; Holmes, J.D. Non-Equilibrium Induction of Tin in Germanium: Towards Direct Bandgap Ge1−xSnx Nanowires. Nat. Commun., 2016, 7, 11405. Chen, W.; Yu, L.; Misra, S.; Fan, Z.; Pareige, P.; Patriarche, G.; Bouchoule, S.; Roca i Cabarrocas, P. Incorporation and Redistribution of Impurities Into Silicon Nanowires During Metal-Particle-Assisted Growth. Nat. Commun., 2014, 5, 4134. Yu, L.; Oudwan, M.; Moustapha, O.; Fortuna, F.; Roca i Cabarrocas, P. Guided Growth of InPlane Silicon Nanowires. Appl. Phys. Lett., 2009, 95, 113106. Yu, L.; Roca i Cabarrocas, P. Growth Mechanism and Dynamics of In-Plane Solid-Liquid-Solid Silicon Nanowires, Phys. Rev. B.; 2010, 81, 085323. Wen, C.Y.; Tersoff, J.; Reuter, M.C.; Stach, E.A,; Ross, F.M. Step-Flow Kinetics in Nanowire Growth. Phys. Rev. Lett., 2010, 105, 195502. Reitano, R.; Smith, P.M.; Aziz, M.J. Solute Trapping of Group III, IV, and V Elements in Silicon by an Aperiodic Stepwise Growth Mechanism. J. Appl. Phys., 1994, 76, 1518-1529. Chen, W.; Yu, L.; Misra, S.; Fan, Z.; Pareige, P.; Patriarche, G.; Bouchoule, S.; Roca i Cabarrocas, P. Incorporation and Redistribution of Impurities Into Silicon Nanowires During Metal-Particle-Assisted Growth. Nat. Commun., 2014, 5, 4134. Misra, S.; Yu, L.; Chen, W.; Roca i Cabarrocas, P. Wetting Layer: The Key Player in PlasmaAssisted Silicon Nanowire Growth Mediated by Tin. J. Phys. Chem. C, 2013, 117, 17786-17790. Nebol'sin, V.A.; Shchetinin, A.A. Role of Surface Energy in the Vapor–Liquid–Solid Growth of Silicon. Inorganic Materials, 2003, 39, 899-903. Chambouleyron, I.; Marques, F.C.; Dionisio, P.H.; Baumvol, I.J.R..; Barrio, R.A.; Mössbauer Study of Hydrogenated Amorphous Germanium‐Tin Thin‐Film Alloys. J. Appl. Phys., 1989, 66, 2083-2090. Chambouleyron, I.; Marques, F.C.; Use of Hydrogenation in The Study of The Properties of Amorphous Germanium Tin Alloys. J. Appl. Phys., 1989, 65, 1591-1597. Kulikov, G.S.; Khodzhaev, K.K.; Diffusion Doping of Undoped Hydrogenated Amorphous Silicon With Tin. Semiconductors, 1998, 32, 123-123. Dash, J. K.; Chen, L.; Lu, T.-M.; Wang, G.-C.; Zhang, L. H.; Kisslinger, K.; Metal-Enhanced Ge1xSnx Alloy Film Growth on Glass Substrates Using a Biaxial CaF2 Buffer Layer. CrystEngComm, 2014, 16, 8794-8804. Rojas-López, M.;Navarro-Contreras, H.; Desjardins, P.; Gurdal,O.; Taylor, N.; Carlsson, J. R. A.; Greene, J. E.; Raman Scattering From Fully Strained Ge1−xSnx (x⩽0.22) Alloys Grown on Ge(001)2×1 by Low-Temperature Molecular Beam Epitaxy. J. Appl. Phys., 1998, 84, 2219-2223. Moutanabbir, O.; Isheim, D.; Blumtritt, H.; Senz, S.; Pippel, E; Seidman, D.N. Colossal Injection of Catalyst Atoms into Silicon Nanowires. Nature, 2013, 496, 78-82. Reitano, R.; Smith, P. M.; Aziz, M.J. Solute Trapping of Group III, IV, and V Elements in Silicon by an Aperiodic Stepwise Growth Mechanism. J. Appl. Phys., 1994, 76, 1518-1529. Aziz, M. J.; Tsao, J.Y.; Thompson, M.O.; Peercy, P.S.; White, C.W. Solute Trapping: Comparison of Theory with Experiment. Phys. Rev. Lett., 1986, 56, 2489-2492. Aziz, M.J. Model For Solute Redistribution During Rapid Solidification. J. Appl. Phys., 1982, 53, 1158-1168. Goldman, L.M.; Aziz, M.J. Aperiodic Stepwise Growth Model For The Velocity and Orientation Dependence of Solute Trapping. J. Mater. Res., 2011, 2, 524-527. 21 ACS Paragon Plus Environment


43 Li, H.; Cui, Y.X.; Wu, K.Y.; Tseng,W.K.; Cheng, H.H.; Chen, H. Strain Relaxation and Sn Segregation in GeSn Epilayers Under Thermal Treatment. Appl. Phys. Lett., 2013, 102, 251907.


Page 22 of 23

Page 23 of 23 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


TOC Graphic


Growth of In-Plane Ge1âxSnx Nanowires with 22 at. % Sn Using a

Recommend Documents

Growth of In-Plane Ge1âxSnx Nanowires with 22 at. % Sn Using a