Vapour-liquid-solid growth of diamond Hexagonal Silicon NanoWires – HexaNW

The main part of the work is based on the comparison between the growth of nanowires experimentally observed at the atomic scale and in real time in NanoMAX (partner LPICM) with theoretical modelling of the growth including thermodynamic calculations (partner C2N) based on ab-initio calculated surface energies (partner LSI). Another important part is the preparation of large samples of 2H nanowires in the PLASFIL plasma-enhanced chemical vapor deposition (PECVD) reactor (LPICM), and their optical characterization (partner IRDEP/IPVF). The most challenging part of the project rests on the in-situ growth in the NanoMAX environmental microscope. More specifically, it is to reproduce the growth conditions of the PLASFIL reactor and especially, the generation of gas radicals, mandatory for the growth of nanowires with Sn catalyst. The solution we have found uses a cyclotron plasma generator, placed into the hydrogen line at the entrance to the TEM column.

The two active axes of work during this first part of the project were WP1 – Growth and WP3 – Modelling. Growth was carried out by PhD students Weixi Wang, funded by the Doctoral school, for studies of 2H nanowires grown in PLASFIL (WP1.1) and Éric Ngo, paid by the project, for the in-situ growth studies in NanoMAX (WP1.2). Experiments in the NanoMAX environmental transmission electron microscope started with the in situ molecular beam epitaxy of Ge NWs, thanks to the collaboration of C2N for the effusion sources. These experiments led to several communications in conferences. Gases for CVD were available in NanoMAX from February 19 on. A first achievement was the growth of Sn-catalysed NWs in situ, thanks to the installation of a cyclotron plasma source in the gas path just at the entrance to the object chamber, to generate the mandatory gas radicals. One of the biggest challenges in the project had thus found a solution. Reproducibly fabricate 2H Si NWs in the PLASFIL PECVD reactor was another challenge. A bronze mixture for the catalyst allowed us to reproducibly obtain 2H structure in a significant part of each deposit. Applying this mixture in NanoMAX allowed us to watch for the first time this month (July 19) the transition between 3C and 2H in a single NW in real time. This is our main result. Regarding Modelling, the active task has been WP3.1 – ab initio calculations of the surface energies of the 2H structure, carried out by Romuald Béjaud, hired by partner LSI as a post doc financed by the project. Romuald calculated most stable structures for whole nanowires of different diameters, with different facets and with or without hydrogen at the Si or Ge surface dangling bounds. A critical radius comes out for the transition from 3C (large nanowires) to 2H (small nanowires): in addition to giving surface energies, the calculations show that the 2H polytype should be more stable than the standard 3C structure in small nanowires.

The most outstanding feature of our studies is the observation in real time of the switching between 3C cubic and 2H hexagonal during in-situ growth in the NanoMAX environmental transmission electron microscope (July 19). This achievement has been attained thanks to the development of the catalyst studies in PLASFIL. It is in PLASFIL that we have started to use very well controlled composition of Sn and Cu catalysts on different substrates (wafer, cantilevers and TEM grids). This has led to a successful increase of the number of 2H nanowires as well as to the successful in-situ observation when the catalysts with optimized composition had been used in NanoMAX. It must also be underlined that the fast and positive results of ab-initio calculations, showing the importance of small size to obtain the 2H structure, have contributed to stimulate the efforts of experimentalists to reproduce in NanoMAX the PLASFIL growth conditions that had produced the smallest wires. For the two remaining years of the project, further observations of the switching from 3C to 2H in NanoMAX are most important tasks. The start of the modelling of the growth (partner C2N), end of this year, will need several of them to draw a general understanding. This understanding, in turn, will be a key help for the final goal of reproducingly fabricate 2H NWs with a high yield. This task will require to continue exploring the growth parameter space, particularly with the PLASFIL reactor. The start of the optical characterisation of the deposits (partner IRDEP/IPVF) will allow us to sort the samples as a function of their properties. It will be important, particularly, to see how multiple twinning and 2H polytypism correlate with optical properties.

Oral communication: Molecular beam epitaxy of germanium in the atomic-resolution transmission electron microscope. 4th International Conference on In Situ and Correlative Electron Microscopy (CISCEM), Sarrebruck, Allemagne, 10-12 octobre 2018. www.ciscem2018.de/program/program/
Publication: É. Ngo, et al., Microsc. Microanal. 25, 47-48 (2019)
Oral: In situ molecular beam epitaxy in the atomic-resolution transmission electron microscope, J.-L. Maurice et al.. Workshop franco-tchèque «Barrande Nano and Advanced Materials Workshop«, École polytechnique (Palaiseau), 18-19 octobre 2018
Poster: In situ TEM study of side-wall effects on the growth of Ge nanowires by the Vapour-Liquid-Solid technique, É. Ngo et al., E-MRS, Nice, France, May 19, www.european-mrs.com/meetings/2019-spring/symposia-program
Oral: Croissance de nanofils de germanium par épitaxie par jets moléculaires in situ dans un microscope électronique en transmission, É. Ngo et al., GdR nanoperando, Lyon, nov. 18 nanoperando.sciencesconf.org/resource/page/id/5
Poster: Effets de surface dans des nanofils de germanium synthétisés in situ dans un microscope électronique en transmission, É. Ngo et al., Colloque de la SFmu juil. 19, Poitiers, colloque.sfmu.fr/fr/programme
Oral: Surface energy calculations of hexagonal silicon nanowires, O. Hardouin Duparc, R. Béjaud, Workshop franco-tchèque «Barrande Nano and Advanced Materials Workshop«, École polytechnique (Palaiseau), 18-19 octobre 2018
Poster: In silico (CD) study of hexagonal diamond (HD) nanowires in silicon, Congrès international iib2019, Paris, 1-5 juillet 2019
Poster: Étude in silico (CD) des surfaces de nanofils en silicium HD, R. Béjaud et O. Hardouin Duparc, Congrès Matériaux 2018, Strasbourg, 19-23 novembre 2018.
Oral: Growth of Si nanowires with cubic and hexagonal structures by PECVD for photovoltaic applications, W. Wang, et al., Congrès Matériaux 2018, Strasbourg, 19-23 novembre 2018.

The diamond hexagonal (2H) phase does not appear in the pressure-temperature phase diagram of silicon. However, the partner Laboratoire de physique des interfaces et des couches minces (LPICM, CNRS, École polytechnique, Université Paris-Saclay) has demonstrated that silicon nanowires having the 2H phase can be produced by the vapour-liquid-solid (VLS) method in a plasma-enhanced chemical vapour deposition (PECVD) reactor. The goal of the HexaNW project is to understand what stabilises that phase during growth in order to establish a protocol to reliably reproduce it. In order to understand that mechanism, we will use the atomic-scale observation of the growth, in situ, in the NanoMAX environmental transmission electron microscope (ETEM), and its modelling by thermodynamic calculations using surface and interface energies that will be ab-initio calculated. NanoMAX is the environmental part of the “EquipeX” TEMPOS.

The calculated band structure of 2H Si makes it a better absorber of the solar spectrum than standard diamond-cubic (3C) silicon (Amato, Nano Lett. 16, 5694, 2016) and its band gap would be direct in nanowires or under stress (Rödl, …, and Guillemoles, Phys. Rev. B 92, 045207, 2015). These electronic properties, together with the natural abundance of the element and its environmental innocuity, make it a wonderful candidate for future nano opto-electronic devices. LPICM develops solar cells based on radial junctions around crystalline nanowires grown by tin-catalysed PECVD. Quite remarkably, the amorphous silicon deposited around a 2H core in such devices would apply a stress that may possibly be used to tune the band gap of the 2H phase.

Although 2H Si does not exist at equilibrium, it has been found in various systems under stress. It has also been obtained deliberately by epitaxy on wurtzite GaP nanowires. But, before our observations, it had never been demonstrated in as grown silicon nanowires without ambiguity. Our observations are uniquely performed in a zone axis of the type [110] (3C)/ [11-20] (2H), where the characterisation of the phase is direct (Tang, Maurice, et al., Nanoscale, 2017). The 2H nanowires we prepare differ from those usually studied by (i) their small size (5 nm), (ii) their low-temperature fabrication by plasma-enhanced VLS and (iii) their liquid tin catalyst.
In addition to understanding the mechanisms of growth, the goal of HexaNW will also be to know the properties that will come out of such objects, as a function of their size and polytype. Its prospects are to develop opto-electronic devices or photovoltaic cells taking advantage of the unique band structure of 2H Si nanowires.

The project will run over a period of 42 months. The programme is the following: growth will be optimised at LPICM in the reactor that already produces the phase. Developing this part will be the task of the PhD student hired by the project. It will be observed in the NanoMAX ETEM in real time and at atomic resolution. Given the risk of this task, the in situ study will also include molecular beam epitaxy of Ge NWs by partner C2N in conditions where the 2H structure would form. Major opto-electronic characterizations of the material will be performed at IRDEP on nanowires grown ex-situ with the help of two interns. Cathodoluminescence (LPICM-C2N) will bring information on the individual luminescence properties. Nucleation and growth will be modelled at C2N (CNRS, Université Paris-Sud, Université Paris-Saclay), with the help of a one-year postdoctoral fellow paid by the project, using surface and interface energies ab-initio calculated at LSI (CEA-CNRS-École polytechnique, Université Paris-Saclay), with the help of a second one-year postdoctoral fellow. The correlation of atomic-scale observations with this modelling will lead to a fine understanding of the ways the Si 2H phase stabilises itself.