Search

Supplemental Material; Doping is a fundamental property of semiconductors and constitutes the basis of modern microelectronic and optoelectronic devices. Their miniaturization requires contactless characterization of doping with nanometer scale resolution. Here, we use low- and room-temperature cathodoluminescence (CL) measurements to analyze p-type and n-type GaAs thin films over a wide range of carrier densities ($2\times 10^{17}$ to $1\times 10^{19}$ cm$^{-3}$). The spectral shift and broadening of CL spectra induced by shallow dopant states and band filling are the signature doping. We fit the whole spectral lineshapes with the generalized Planck's law and refined absorption models to extract the bandgap narrowing (BGN) and the band tail for both doping types, and the electron Fermi level for n doping. This work provides a rigorous method for the quantitative assessment of p-type and n-type carrier density using CL. Taking advantage of the high spatial resolution of CL, it can be used to map the doping in GaAs nanostructures, and it could be extended to other semiconductor materials.

International audience; Since LaPierre's first theorization of a nanowires (NWs)-based tandem solar cell (TSC) on Si substrate, 1 other theoretical studies have been carried out to investigate the high potentialities of such devices, 2 which can virtually reach an efficiency around 34 %. In spite of these very promising results, the research in this field has not yet developed that much on a practical level: to the best of our knowledge only one prototype of TSC has been realized so far by Yao et al, 3 and although not completely optimized, it reached the high conversion efficiency value of 11.4 %. Based on these considerations, we aimed to develop a fully optimized TSC capable to achieve higher efficiencies. 4 This consists in a top cell made of an ordered array of GaAs/AlGaAs core/shell NWs, with a radial p-in junction in the AlGaAs shell, directly grown on a bottom cell of Si(111), as shown in Figure 1(a). The optimization work has been carried out in different steps, i.e. theoretical simulations to identify the best set up for the array, optimization of the NW structure, composition and morphology, investigation for the most suitable passivation shell and growth on patterned substrates. The last point has notably lead to significant results, since we develop a protocol to obtain very high vertical yields (from 80 % to 90 %) of GaAs/AlGaAs core/shell NWs on large patterned areas (0.9 cm x 0.9 cm), as shown in Figure 1(b). Following these promising results, we are currently committed to using our ordered arrays of NWs for the fabrication of a TSC, now focusing our work on encapsulating and contacting the nanostructures: despite the complexity of the system, EBIC results in this respect (cf. V. Piazza's abstract) are encouraging and suggest that the realization of the TSC is at our fingertips. 5 Figure 1. (a) Scheme of the ultimate structure of the TSC. (b) Ordered array of p-GaAs core / p-in AlGaAs shell NWs on Si(111) with a high vertical yield (80-90 %); white scale bar corresponding to 2 µm.

We report preliminary results on InGaP/InGaAs/Ge photovoltaic cells for concentrated terrestrial applications, monolithically integrated on engineered Si(001) substrates. Cells deposited on planar Ge/Si(001) epilayers, grown by plasma-enhanced chemical vapor deposition, provide good efficiency and spectral response, despite the small thickness of the Ge epilayers and a threading dislocation density as large as 107/cm2. The presence of microcracks generated by the thermal misfit is compensated by a dense collection grid that avoids insulated areas. In order to avoid the excessive shadowing introduced by the use of a dense grid, the crack density needs to be lowered. Here, we show that deep patterning of the Si substrate in blocks can be an option, provided that a continuous Ge layer is formed at the top, and it is suitably planarized before the metalorganic chemical vapor deposition. The crack density is effectively decreased, despite that the efficiency is also lowered with respect to unpatterned devices. The reasons of this efficiency reduction are discussed, and a strategy for improvement is proposed and explored. Full morphological analysis of the coalesced Ge blocks is reported, and the final devices are tested under concentrated AM1.5D spectrum. Copyright © 2016 John Wiley & Sons, Ltd.

Our work, focused on the heteroepitaxial Ge/Si (001) system shows that patterned Si substrates in appropriate pillar arrays, featuring micrometric dimensions within the common deep-etching capabilities and array size suitable for several applications (up to few hundreds microns), do provide exceptional compliance to the thermal strain of a continuous Ge film deposited on top. Micro-Raman and XRD measurements on the Ge suspended layer, as obtained by Low Energy Plasma Enhanced Chemical Vapor Deposition, confirm the theoretical predictions. Actually, the sharp onset with pillar aspect ratio of such an effect is demonstrated to be triggered by the free rotation of the pillars, as induced by the thermal contraction of the film. The strain relaxation is deeply analyzed both by FEM simulations and experimental XRD results, showing a very good quantitative agreement. Our findings have far reaching consequences beyond the crystalline heteroepitaxial systems, whenever a film of any microstructure, or composition, is deposited in suspended patches on suitable pillar arrays.

The move from dimensional to functional scaling in microelectronics has recently revamped the interest towards integration of high-quality Ge on Si [1]. However, due to the 4% lattice misfit, this is quite a technological challenge. In ref. [2] we showed the possibility to grow vertically-aligned micrometric Ge crystals, separated by nanometric gaps, on top of Si pillar patterned substrates, without dislocations at the top region. Herein we show by simulation-driven experiments, how these structures can evolve into a suspended Ge film by simply exploiting their temperature-driven coalescence. By closely comparing experimental data and simulation results based on a phase-field model of surface diffusion [3], we demonstrate that the massive morphological evolution is essentially driven by the lowering of surface-curvature gradients. Scanning Tunneling Electron Microscopy (STEM) and High Resolution - Transmission Electron Microscopy (HR-TEM) measurements indicate that the resulting film has very high crystal quality. In particular, dislocations are not observed to nucleate when the crystals start to coalesce, as confirmed by etch-pit analysis. Our approach thus provides a viable path for the integration on Si of Ge layers with very low defect density. [1] R. Pillarisetty, Nature 479, 324 (2011); [2] C.V. Falub et al., Science 335, 1330 (2012); [3] M. Salvalaglio, R. Backofen, R. Bergamaschini, F. Montalenti, A. Voigt, Cryst. Growth Des, in press (doi: 10.1021/acs.cgd.5b00165)

Droplet epitaxy (DE) is a growth technique based on molecular beam epitaxy proposed in 1991 by Dr. N. Koguchi [1]. This method allows for the fabrication of wetting layer free, lattice-matched and strain-free III-V nanostructures self-assembled, reducing the size dispersion between 5 and 20%. Thanks to the separation of the group III and group V atom irradiation on the substrate, it is possible to obtain a fine tuning of the quantum dot (QD) shape and density by simply changing the parameters that control the group III atom diffusion on the surface. It is possible to select a density between 107 and 1011 cm-2 and the aspect ratio of the dots, thus allowing to change the quantum confinement and to tailor the optical and electronic property of the dots to fit the needs of different applications. In this presentation we discuss the main aspects of the growth and characterization of DE nanostructures, together with the advantage of shape control for different applications recently developed. The first application is related to the fabrication of intermediate band solar cells, introducing a layer of quantum dots grown by DE in a standard AlGaAs solar cell. By tuning the size of the QDs it is possible to change the position of the intermediate band, and by tuning the aspect ratio of the QDs the high energy states of the QDs can also be tuned in order to have a small electron-phonon coupling with the barrier. Moreover, the lack of defect and wetting layer states can greatly reduce thermal escape of carriers from the intermediate band, leaving photon-induced transitions the dominant ones, as requested by intermediate band theory [2]. The second application is related to the fabrication of GaAs single photon emitters integrated on Si substrates. Hanbury-Brown-Twiss measurements demonstrates how the single nanostructure is a single photon emitter up to temperature of liquid nitrogen, thus demonstrating how DE makes also possible the growth of bright III-V quantum emitters on silicon substrates and paving the route to the integration of optically efficient III-V nanostructures on CMOS technology [3]. The last application is the growth of ultra high density QDs for the fabrication of quantum dot infrared photodetectors (QDIP) active in the range between 2 and 8 µm, as a first step toward the the fabrication of a III-V based QDIP integrated on Si [4].[1] N.Koguchi, S.Takahashi, T.Chikyow, Journal of Crystal Growth 111, 688 (1991) [2] A.Scaccabarozzi, S.Adorno, S.Bietti, M.Acciarri, S.Sanguinetti, Physica Status Solidi (RRL) - Rapid Research Letters 3, 173, (2013) [3] L.Cavigli, S.Bietti, M.Accanto et al., Applied Physics Letters, 100, 231112, 2012. [4] J.Frigerio, G.Isella, S.Bietti, S.Sanguinetti, SPIE Newsroom, 2013

This thesis presents my Ph.D. work about quantum dot GaAs/AlGaAs solar cells grown by droplet epitaxy, exploring the potential of this materials system for the realization of intermediate band photovoltaic devices. In the first chapter a general introduction to the field of solar energy is given, outlining the reasons why this research has been performed. The physics of the photovoltaic cell is briefly explained in its most important points, to give the reader clear understanding of what is presented in the following chapters. Intermediate band devices are presented in the second chapter. The theoretical foundations presented do not aim at constituting an exhaustive explanation of the theory underlying intermediate band solar cells, but the scope is again to give clear understanding of the characterization of the quantum dot devices reported in the following chapters. A survey of the state of the art in the field is given, pointing out the differences with our technology. The initial part of my Ph.D. work was spent in developing the technology to design and grow (Al)GaAs photovoltaic devices, as well as the characterization techniques required to understand the behavior of such devices. In chapter 3 the method developed to design the solar cell structure is illustrated, and in chapter 5 the experimental setup used for characterization is presented, along with the measurements on the single junction devices realized during this work. Chapter 4 is dedicated to the description of the growth and fabrication methods used to grow the samples reported here. The development of the fabrication technology proceeded in close contact with the characterizations of the devices, in order to optimize the process. Finally in chapter 6 the results on quantum dot photovoltaic cells are reported: the key working principles of intermediate band devices have been demonstrated with our materials system, and this, to the knowledge of the author, is the first time that strain free quantum dot solar cells are reported of intermediate band behavior. The role of defects in the AlGaAs matrix is explained in connection with both the optical and electrical characterizations presented.

Quantum dot intermediate band (IB) solar cells have been proposed in order to increase the efficiency of traditional solar cells and immediately attracted a lot of attention and research since the time they were first proposed [1]. The IB introduces an extension of the absorption coefficient of the semiconductor to lower energies, allowing a more thorough collection of the solar spectrum, via a two-step absorption of low-energy photons. Quantum dot (QD) structures are good candidates for IB solar cells, because their confined energy levels can overlap and form a miniband in dense arrays and the position of these levels and bands can be tuned varying the size and spacing of the QDs. The IB working mechanisms have been demonstrated for InAs QDs in GaAs [2], and a lot of research is devoted to reduce the problems due to strain, and defect nucleation increasing carrier escape [3]. Droplet epitaxy (DE) [4] is a molecular beam epitaxy technique that allows for the growth of quantum dots of materials lattice matched to the barrier and the removal of the wetting layer. DE makes possible to indipendently control density, size and shape of the nanostructures. Densities as high as some 1011 cm-2 per layer have been reported [5] and potentially a large number of layers can be stacked because the system is strain-free, leading to a much higher density of states in the IB than the more conventional Stranski-Krastanov techniques. Since it is a strain-free technique, there are virtually no defects in DE-grown materials, and this is fundamental to have high performance devices. Moreover, DE nanostructures can be grown without the presence of a wetting layer [6], that would introduce unwanted quantum-well-like states in the system. By tuning the size of the QDs it is obviously possible to change the position of the IB, and by tuning their aspect ratio the high energy states of the QDs can also be tuned in order to have a small electron-phonon coupling with the barrier. The lack of defect and wetting layer states can greatly reduce thermal escape of carriers from the IB, leaving photon-induced transitions the dominant ones, as requested by IB theory [1]. For these reasons DE is a good candidate for the realization of QD-IB solar cells. We demonstrate the key working principle of IB materials, that is the production of sub-gap two-photon photocurrent, with Al0.3Ga0.7As solar cells containing GaAs QDs grown by DE (figure 1a). The devices were illuminated at 15K by continuous monochromatic light and chopped broadband IR light (1.5 – 20 m), and the signal was demodulated by a lock-in amplifier. The continuous light pumps electrons into the QDs and the chopped IR promotes them to the CB where they can be collected generating an electric signal. As a control, a reference sample without QDs was grown and measured in the same conditions, but no two-photon signal was detected (see figure 2b). Measurements show that the shape of the two-photon photocurrent signal is different from the VB-CB photoresponse, indicating clearly that a two-photon process involving the QD levels is taking place. The two-photon signal response is in good agreement with the PL spectrum of the QD samples (figure 1b), that lets us easily understand the position of the energy levels of the system. [1] A. Luque, A. Martì, Physical Review Letters, 78, 5014 (1997). [2] A. Martì, E. Antolìn, C. R. Stanley, C. D. Farmer, N. Lòpez, P. Dìaz, E. Cànovas, P. G. Linares and A. Luque, Physical Review Letters 97, 247701 (2006). [3] A. Luque and A. Martì, Prog. Photovolt: Res. Appl. 9, 73-86 (2001) [4] N. Koguchi and K. Ishige, Japanese Journal of Applied Physics 32, 2052–2058 (1993). [5] M. Jo, T. Mano, Y. Sakuma and K. Sakoda, Applied Physics Letters 100, 212113 (2012) [6] S. Sanguinetti, K. Watanabe, T. Tateno, M. Wakaki, N. Koguchi, T. Kuroda, F. Minami, and M. Gurioli, Applied Physics Letters 81, 613 (2002)

Quantum dot intermediate band (IB) solar cells have been proposed in order to increase the efficiency of traditional solar cells and immediately attracted a lot of attention and research since the time they were first proposed. The IB introduces an extension of the absorption coefficient of the semiconductor to lower energies, allowing a more thorough collection of the solar spectrum, via a two-step absorption of low-energy photons. Quantum dot (QD) structures are good candidates for IB solar cells, because their confined energy levels can overlap and form a miniband in dense arrays and the position of these levels and bands can be tuned varying the size and spacing of the QDs. The IB working mechanisms have been demonstrated for InAs QDs in GaAs, and many efforts are devoted to reduce the problems due to strain, and defect nucleation increasing carrier escape. Droplet epitaxy (DE) is a molecular beam epitaxy technique that allows for the growth of quantum dots of materials lattice matched to the barrier and the removal of the wetting layer. DE makes possible to indipendently control density, size and shape of the nanostructures. Densities as high as some 1011 cm-2 per layer have been reported and potentially a large number of layers can be stacked because the system is strain-free. Being a strain-free technique, there are virtually no defects in DE-grown materials. By tuning the size of the QDs it is obviously possible to change the position of the IB, and by tuning their aspect ratio the high energy states of the QDs can also be tuned in order to have a small electron-phonon coupling with the barrier. The lack of defect and wetting layer states can greatly reduce thermal escape of carriers from the IB, leaving photon-induced transitions the dominant ones, as requested by IB theory. For these reasons DE is a good candidate for the realization of QD-IB solar cells. We demonstrate the key working principle of IB materials, that is the production of sub-gap two-photon photocurrent, with Al0.3Ga0.7As solar cells containing GaAs QDs grown by DE. Measurements show that the shape of the two-photon photocurrent signal is different from the VB-CB photoresponse, indicating clearly that a two-photon process involving the QD levels is taking place. The two-photon signal response is in good agreement with the PL spectrum of the QD samples, that lets us easily understand the position of the energy levels of the system.

We report the fabrication procedure and the characterization of an Al0.3Ga0.7As solar cell containing high-density GaAs strain-free quantum dots grown by droplet epitaxy. The production of photocurrent when two sub-bandgap energy photons are absorbed simultaneously is demonstrated. The high quality of the quantum dot/barrier pair, allowed by the high quality of nanostructured strain-free materials, opens new opportunities for quantum dot based solar cells. Left: Scheme of the fabricated quantum dot solar cell. Right: 1 × 1 μm2 image of the dot layer. Copyright © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

We have recently shown that crystal defects and internal stresses commonly hampering Ge heteroepitaxy on Si can be greatly reduced by promoting the nucleation of micro-crystals at the top of Si pillars deeply patterned on the substrate [1]. Ge-on-Si epitaxy by low-energy plasma-enhanced CVD (LEPECVD), allows for controlling the microcrystal faceting, in turn affecting the evolution of threading dislocations and providing complete expulsion of threading arms to the microcrystal sidewalls [2]. In addition, the microcrystals are self-assembled in dense arrays, with a spacing ranging from tens to one hundred nanometers. The growth mechanism of the micro-crystals is an intriguing “nano” effect, which has been interpreted in terms of independent facet growth. This effect is not a unique feature of Ge deposition by LEPECVD, but a more general growth mode, which can be extended to other materials systems and deposition techniques. GaAs microcrystals deposited by MBE on patterned Si substrates show as well the complete tessellation of the surface by closely separated microcrystals. The complete elimination of thermal strain and reduction of threading dislocations is confirmed by the high photoluminescence yield. [1] C. V Falub et al. Science 2012, 335, 1330–4 ; [2] A. Marzegalli et al., Advanced materials 2013, 25, 4408