Electronic & optical nature of silicon nanostructures: doping, interface effects & strain

Resume : Silicon quantum dots (Si QDs) are a promising alternative to toxic and rare material QDs that are researched or used in optoelectronics, photonics and bio-imaging. For competitive emission properties, however, the radiative rate needs to be considerably improved. This is the case in Si QDs capped with organic ligands; Nevertheless, despite high radiative rates approaching those of direct bandgap QDs [1,2], the quantum yield (QY) in the visible range remains comparatively low (<20%) [1,2]. Improved understanding of the processes underlying the ensemble photoluminescence (PL) and QY of these materials can be achieved from single QD spectroscopy, allowing study of the properties of individual emitters that are otherwise obscured in ensemble measurements. As generally in all quantum emitters, the QY is critically influenced by PL blinking. In addition, poor surface passivation can further reduce the QY by introducing competing non-radiative channels. In both mechanisms surface states play a major role. By single QD spectroscopy, we examine the effect of ligand type on the blinking of colloidal Si QDs and moreover, we study the competition of the radiative and non-radiative recombination pathways by decoupling them by means of a Drexhage-type experiment. Understanding the role of surface will help us to improve the QY of Si QDs for application in lighting technologies. [1] K. Dohnalova et al., Light: Science and Applications 2 (2013) e47 [2] K. Kusova et al., ACS Nano 4 (2010)

Resume : Silicon nanocrystals (Si NCs) embedded in SiO2 have been extensively studied because of their potential to fabricate novel optoelectronic and photonic devices. Their optical properties are strongly affected by both their size and local environment. The aim of the present study is to investigate the role of the exerted pressure over the NCs on their optical emission, by means of optical characterization under high hydrostatic pressure. The samples consist of Si NC/SiO2 superlattices containing NCs with average crystalline diameters of either 4.1 or 3.2 nm. High-pressure Raman scattering and photoluminescence (PL) measurements were performed up to 5.5 GPa with a membrane-type diamond anvil cell, using the 532-nm line of a Nd:YAG laser. The Raman spectra of the two samples allowed determining the phonon pressure coefficient, yielding a value of (8.5 ± 0.3) cm-1/GPa in both samples, which is much larger than the one for bulk Si (5.1 cm-1/GPa). This result is ascribed to a strong pressure amplification effect due to the larger compressibility of the SiO2 matrix. The PL spectra exhibit a lower energy band that barely depends on pressure, which was ascribed to defects. The pressure coefficients of the higher energy band were found to be much larger than the one for bulk Si, with values very similar for both Si NC sizes: (-35 ± 8) and (-27 ± 6) meV/GPa for 4.1 and 3.2 nm, respectively. The analysis of the PL data, once the pressure amplification is included, revealed that the emission from Si NCs is consistent with that of the indirect transition of Si.

Resume : Silicon nanoparticles potentially enable new applications in silicon based optoelectronics and photovoltaics. Single layers of nitrogen free Si rich oxides with varying Si excess concentrations were deposited by PECVD [1,2] followed by a subsequent high temperature annealing to induce phase separation. The structural properties of the layers are investigated by plane view TEM using high resolution as well as energy filtered TEM. At low Si excess, isolated single crystalline Si particles with low density are observed indicating an incomplete phase separation. An increase of Si excess leads to larger Si particles along with an increased intrinsic crystal defect density. Furthermore, we find a transition point near SiOx=0.4, where the isolated Si particles merge into a continuous network. In addition we study the influence of annealing conditions and find negligible differences that are indicative of fast phase separation on the timescale of seconds and significantly slower pa rticle growth due diffusion processes. The importance of the results is discussed in the context of transport properties that affect a possible Si nanocrystal device performance. [1] Laube et al. JAP, 116, 223501, 2014 [2] Gutsch et al. BJnano 6, 964, 2015

Resume : Quantum confinement strongly affects properties of hot carriers in nanocrystals. In particular their cooling rates and thus effective lifetime are influenced by reduction of phonon emission and energy recycling through a combination of enhanced Auger recombination and impact excitation. This, among others, promotes enabling multiple exciton generation rates and energy transfers to the outside of the nanostructure. In my presentation, I will discuss some results of investigations of these effects in Si and Ge nanocrystals embedded in SiO2: 1. Carrier multiplication in Ge nanocrystals: I will report on the recent identification of this effect in materials prepared by co-sputtering 2. Carrier multiplication in Si nanocrystals. Here I will discuss the on-going investigations of spectral modification of emission induced by the carrier multiplication process providing a unique spectral fingerprint of this special process. 3. Energy recycling processes for hot carriers in Si nanocrystals, enabled by interplay between impact excitation and Auger recombination of multiple excitons in the same nanocrystal, leading to effective increase of free electron temperature 4. Excitation of Er emitters by hot carriers optically generated in Si nanocrystals, demonstrating that this energy transfer path can successfully compete with phonon emission. 5. Energy diffusion within an ensemble of Si nanocrystals, leading to peculiarities in exciton characteristics. [1] M.T. Trinh et al., Nature Photonics 6, 316-321 (2012). [2] S. Saeed et al., Nature Communications 5:4665 (2014) [3] F. Priolo, T. Gregorkiewicz, M. Galli, and T. Krauss, Nature Nanotechnology 9, 19 (2014) [4] S. Saeed et al., NPG Light: Science and Applications, 4, e251 (2015)

Resume : Photo- and electro-luminescence quantum yield (QY) is a crucial parameter for materials to be used in light-emitting devices, fluorescent probes etc. We have developed a compact set-up and procedure for reliable measurements of QY over a broad excitation spectral range which uses excitation by LEDs and a carefully calibrated spectroscope [1] and applied them to various nanomaterials, mostly Si nanocrystals (SiNCs). Analysis of the QY spectral dependence can reveal important information on non-radiative recombination channels, carrier multiplication processes, collective phenomena in dense ensembles of NCs etc. [2]. In this contribution we summarize our experiments on various forms of SiNCs, namely multilayers of Si nanocrystals (NCs) separated by SiO2 barriers and doped SiNCs in the form of liquids and deposited layers. Multilayer samples enabled us to clearly demonstrate that the increase of barrier thickness from about 1 to larger than 2 nm induces doubling of the PL QY value which corresponds to the change of number of close neighbours in the hcp structure [3]. The temperature dependence of PL QY suggests that the PL QY changes are due to a thermally activated transport of excitation into non-radiative centers in dark NCs or in the matrix. Finally, we shall discuss all possible phenomena that contribute to limiting PL QY values in ensembles of SiNCs. [1] J. Valenta, Nanoscience Methods 3 (2014) 11-27 (OA). [2] D. Timmerman et al. , Nature Nanotechnology 6 (2011) 710. [3] J. Valenta et al., Appl. Phys. Lett. 105 (2014) 243107.

Resume : The absolute photoluminescence (PL) quantum yield (QY) of multilayers of Si nanocrystals (NCs) separated by SiO2 barriers were thoroughly studied as function of the barrier thickness, excitation wavelength, and temperature. The temperature dependence of PL QY suggests that the PL QY changes are due to a thermally activated transport of excitation into non-radiative centers in dark NCs or in the matrix in agrrement with previous results obtained from time-resolved PL spectroscopy. The PL QY excitation spectra show no significant changes upon changing the barrier thickness and no clear carrier multiplication effects. Moreover, we find that there is no PL spectral shift as a function of excitation energy clearly evidencing the absence of space-separated quantum cutting effects. [1] Valenta et al., Appl. Phys. Lett. 105, 243107 (2014)

Resume : A monosilane (SiH4) and oxygen (O2) based plasma-enhanced chemical vapor deposition process (PECVD) for the growth of silicon-rich oxide / silicon dioxide superlattices [1] was developed. In contrast to the conventional nitrous oxide (N2O) based oxynitride-process [2], we achieved thereby PECVD-grown size-controlled silicon nanocrystals in pure, N-free silicon dioxide matrix. We present a detailed study based on optical (PL) and electrical (C-V, I-V) measurements that reveal the different properties of nominally identical Si nanocrystals in oxynitride and N-free oxide matrix. Most strikingly we find negligible differences in the optical performance (PL quantum yield), whereas substantial differences in the current transport and transient charging behaviour persist. The role of the pure oxide vs. oxinitride matrix on the properties of Si quantum dots is discussed in the context of potential applications in photovoltaics and optoelectronics. [1] M. Zacharias et al., APL 80, 2002 [2] Laube et al. JAP, 116, 223501 [3] S. Gutsch et al., JAP 113, 2013

Resume : Silicon Nanoparticles (Si NPs) have been produced through a high throughput, low-temperature plasma synthesis which yields faceted, crystalline and monodisperse NPs ~ 100 nm in size. When deposited on a semiconductor substrate, these nanostructures have shown potential towards the trapping of light within a single NP, where the unique faceted nature of the NPs was a key factor in enhancing the photoluminescence yield of the substrate by increasing the optical path length of the impinging photons [1]. Applying this idea, we perform the assembly of the silicon NPs into large-scale superstructures, and study the resulting modifications in optical properties. This has been pursued twofold: on a substrate (2-dimensional film) and in solution (3-dimensional structure). In the latter case, the colloids have been brought together via the Critical Casimir effect [2]. The results of this investigation may be attractive for novel applications in the field of photovoltaics, possibly giving birth to a new class of thin-film silicon solar cells. While absorption qualities of bulk silicon are maintained due to large size of the Si NPs, light path is considerably enhanced and layers can be deposited by low-cost low-temperature self-assembly approaches. Furthermore, we expect that faceted Si NPs in the high quality assembly (side-to-side) will ensure high level carrier mobility. [1] Mannino et al., Sci. Rep. 5, 8354 (2015). [2] Nguyen et al., Nat. Commun. 4, 1584 (2013).

Resume : We report on phosphorous (P) doping of Si nanocrystal (SiNC)/SiO2 systems [1]. Relevant P configurations within SiNCs, at SiNC surfaces, within the sub-oxide interface shell and in the SiO2 matrix were evaluated by hybrid density functional theory (h-DFT). Atom probe tomography (APT) and its statistical evaluation provide detailed spatial P distributions. We obtain ionisation states of P atoms in SiNC/SiO2 systems at room temperature using X-ray absorption near edge structure (XANES) spectroscopy in fluorescence yield mode. This non-destructive volume characterization technique leaves the material unchanged which is paramount to probe P in its original state. P K shell energies were confirmed by h-DFT. We found that P diffuses into SiNCs and resides virtually exclusive on interstitial sites. XANES and h-DFT delivered evidence that the ionization probability of P in the SiO2/SiNC system is extremely low; free localized electrons to SiNCs are not provided. [1] D. König, S. Gutsch, H. Gnaser, et al., Sci. Rep. (Nature), Vol. 5, 09702 (2015), DOI: 10.1038/srep09702

Resume : We model fully OH- and NH2-terminated silicon nanocrystals (SiNCs) - Si35(OH)36 and Si35(NH2)36 - by hybrid-DFT [1,2]. We substitute OH [NH2] groups by double-bonded (=) O [=NH] segments and corner Si(OH)2 [Si(NH2)2] segments for bridge-bonded (>) O [>NH] groups. Correlating ground state (GS) gaps and charge transfers from SiNCs to anion groups with atomic ratios of Si to O [N] atoms, we show the specific impact of = and > anion bonds. Excited state (ES) calculations yielded the three lowest singlet transition energies (ES gap, absorption) and oscillator strengths (optical transition probabilities P_ot). ES energies closely follow the trend of GS gaps for OH-terminated SiNCs with =O, while those for SiNCs with >O show an increasing deviation from GS gaps to lower energies with rising numbers of >O. An increased modulation of charge transfer occurs due to low negative ionization of >O atoms which appears to distort the exciton field and thus increases exciton binding energies. For SiNCs with =O atoms, stronger SiNC ionization seems to improve carrier separation, decreasing exciton binding energies and lifting ES gaps. Thus, P_ot of SiNCs rise hyperlinear with the number of =O atoms, while a saturation occurs for SiNCs with >O atoms. Similar results were found for NH2-terminated SiNCs with =NH vs. >NH, with less influence on the electronic and optical SiNC behaviour due to lower SiNC ionization. [1] Phys. Rev. B 78, 035339 (2008) [2] Adv. Mater. Interf. 1, 201400359 (2014)

Resume : Thanks to their specific electronic properties, silicon nanostructures (nanowires, nanoribbons) offer great potential for high integration electronics, and for highly sensitive sensors. In our work, silicon nanostructures are synthesized following classical top-down approach using conventional UV lithography technique fully compatible with the existing silicon technology. Silicon nanowires are fabricated following the spacer method [1]. In this method, amorphous silicon layer, crystallized by thermal annealing, is deposited on a SiO2 step. An anisotropic reactive ion etching of the polycrystalline silicon layer allows sidewall spacers formation used as nanowires. Nanoribbons are made of very thin (< 50nm) polycrystalline silicon layer. However, in both cases, such silicon nanostructures show poor electrical properties due to their crystalline quality. One solution studied here to improve it is the metal induced lateral crystallization method, using a metallic catalyst to create oriented crystals. N channel silicon nanostructures based field effect transistors have been fabricated and tested with top-gate, bottom-gate and gate-all-around architectures. Silicon nanostructure/SiO2 interfaces are studied from electrical characteristics by the determination of the density of states using the incremental method of Suzuki [2]. [1] F. Demami, L. Pichon, R. Rogel, and A. C. Salaun, Mat. Sc. Eng. , 012014 (2009) [2] T. Suzuki, Y. Osaka and M. Hirose, Jap. J. Appl. Phys., vol.21 (1982)

Resume : One of the major challenges toward scaling of microelectronics devices to the nanoscale is the control of the semiconductor's electronic properties. For this aim, the incorporation of dopants with atomic accuracy alters the semiconductor's electronic structure by creating excess carriers. Unfortunately, conventional doping technologies, such as ion implantation and solid source diffusion processes, do not work at such small scales and for non-planar geometries. An alternative strategy to achieve high-quality doping profiles with areal uniformity at the nanoscale is the so-called monolayer doping. Such approach relies on a self-limiting surface reaction that leads to the formation of a monolayer of dopant-containing molecules bonded to the substrate. To date, monolayers containing different dopants have been formed on semiconductor substrate, but the process of monolayer formation has not been entirely characterized, highlighting many open questions. In this work, a systematic study of P-monolayer doping process on SiO2 substrate is presented. A proper experimental protocol is proposed to reduce the amount of contaminants introduced in the sample during the process and to obtain a P delta layer embedded in a SiO2 matrix. A multi-technique (TOF SIMS and RBS) approach allows monitoring the amount of dopants bonded to the substrate and the reproducibility of the process. The quantity of dopant is tuned by changing the processing parameters.

Resume : Si-based efficient light sources, amplifiers, lasers and photodetectors are continuously sought for realization of the on-chip Si-photonics. The major showstopper, however, is the inherent indirect bandgap of the bulk Si. Two groups have recently reported occurrence of direct bandgap transitions in C-capped Si quantum dots (C-SiQDs) [1-3] – both simulated theoretically and confirmed experimentally. While our concept proposes occurrence of such transitions as a result of modified electronic density in the C-SiQD core resulting from the electronegative capping [1,2], the other group suggests that crucial role plays tensile strain induced by the C-linked ligands [3]. Despite non-unified model, experimental evidence of both groups shown dramatically enhanced radiative rate and size-tunable emission [1,3-5]. To better understand the separate role of the field, we will compare two model situations [2]: (i) covalently attached electronegative ligand (C-SiQDs) and (ii) H-capped SiQDs in locally electronegative field – a system, where only external local electrostatic field plays role and no ligand-induced strain can be present. [1] K. Dohnalova, et al., Light: Science & Applications 2 (2013) e47 [2] A. N. Poddubny, K. Dohnalova, Phys. Rev. B 90 (2014) 245439 [3] K. Kusova et al., Adv. Mater. Int. 1 (2014) 1300042 [4] K. Kusova et al., ACS Nano 4 (2010) 4495 [5] K. Dohnalova et al., Small 8 (2012) 3185

Resume : Reliable and precise modelling of electronic and optical properties of semicondcutors heterostructures is crucial for the optimization of low dimensional electronic and optical devices. It is well known that quantum mechanical calculations electronic properties of heterostructures usually encounters computational and/or conceptual difficulties. Over the years various simplified models have been proposed in attempting to achieve the underlying physics of the interface formation and reveal new insights about the chemical trends. In this work, we propose to use an sp3 bond orbital tight binding theory [1,2] to determine the conduction and valence band energy levels and band offsets in heterostructures.. The model considers the nonorthogonality of hybrids of adjacent atoms and the interaction of bonding and antibonding states and metallic contribution to determine the condcution and valence band energies of semiconductors at high symmetry points (e.g., , L and X). The tight binding eigen-energies of given atoms (anion or cation) are obtained from the Hartree-Fock free atom values, including the Uss, Usp and Upp Coluomb corrections in the solid. The valence band offsets are then obtained from the difference between the valence band energies, screened by the optical dielectric constants of constituent semiconductors. The proposed bond orbital sp3 tight binding model considers the interactions between the sp3 hybrids at adjacent sites, with special consideration given to the overlapping of these hybrids such that it cannot be absorbed into a rescaling . This approach also allows one to determine the polarity of the bond and in turn the effective charge, bulk modulus, elastic constants, deformation potential of both valence maximum and conduction band edges at high symmetry points within the Brillouin zone of tetrahedral semiconductor energy band structure in a self-consistent manner. With such approach we provide more reliable predictions for the properties of the electronic band structure of Si1-x-yGexCy , Si1-xGex and Si1-yCy alloys, including the strain-compensated case, from the interpolation between compressive strained Si1-xGex and tensile strained Si1-yCy.

Resume : Manufacturing of highly effective silicon-based light-emitting diodes is a very important technological problem. It is possible to develop a silicon with photoluminescent properties by the irradiation process. Irradiation causes the formation of various defects in silicon structure, including point defects, defect clusters and extended complexes. Until this day there is no detailed experimental data on what types of defects are possible in silicon. Some experimental observations point out the possibility of self-assembly of defects (vacancies and self-interstitials) in silicon during irradiation. In this work theoretical investigation of the stability of various defect complexes including vacancies, self-interstitials and their clusters was performed by means of ab initio as well as molecular dynamic calculations. Ab initio calculations in the framework of density functional theory of silicon crystal with various defects were carried out. Meta-stable defect configurations were found. The software for the molecular dynamic calculations and for the visualization of calculation results was developed. The results of the first-principles calculations were used for the parametrical fitting of Tersoff potential of silicon. The potential was used to perform molecular dynamic calculations of the various configurations of defects in silicon. This approach allowed us to study the dynamic of the atomic system over time and temperature range.

Resume : The current through dielectric embedded Silicon Nanocrystal thin films is strongly dependent on the applied electric field. Here, we demonstrate that this strong dependence can be explained by a black box model based on a space-charge limited current through insulators in the presence of traps. The result render commonly used models to fit the data superfluous.

Resume : Silicon nanocrystals (SiNCs) are a material which may possess photoluminescence (PL) in the visible spectral range. In the case of oxide-passivated SiNCs, the PL is typically composed of a slow-decaying red–orange band (S-band) and of a fast-decaying blue–green band (F-band). Particularly the origin of the F-band is still not fully understood. In order to investigate the F-band of oxide-passivated free-standing SiNC, we have performed an experimental study which combine temperature (from 4K to 300K), temporal (with picosecond resolution) and spectrally-resolved luminescence spectroscopy. This study shows that that the F-band red-shifts only by 35 meV with increasing temperature, which is almost 6 times less than the red-shift of the S-band in a similar temperature range [Ondic et al, Nanoscale 6, 3837 (2014)]. In addition, the F-band characteristic decay time obtained from a stretched-exponential fit decreases only slightly with increasing temperature. These data suggest that the F-band arises from the core-related quasi-direct radiative recombination governed by slowly thermalizing photoholes.

Resume : The dependence of the linear and nonlinear optical properties of silicon nanowires (SiNWs) on their structural properties was investigated. SiNWs with diameter between 40 and 200 nm were grown on crystalline silicon (c-Si) substrates by metal-assisted chemical etching and had a strong scattering of light in the visible and infrared spectral range. Nonmonotonic dependence of the total reflectance of light from SiNWs on their length was discovered. It was found that the indicatrix of the elastic scattering of light with a wavelength of 1064 nm in the SiNW ensembles longer than 2 μm can be approximated by the Lambert's cosine law and the intensity of the scattered light in the rear hemisphere grows logarithmically with increasing the length of SiNWs. It was found that in SiNWs the efficiency of frequency conversion of optical radiation, such as Raman scattering, coherent anti-Stokes Raman scattering and third-harmonic generation, were increased compared to c-Si substrates. Cross-correlation functions of photons scattered in SiNW arrays were measured for the first time and it was found that the interaction of light with SiNWs was multiple times increased compared to c-Si substrates. Nonmonotonic dependence of the intensity of the interband photoluminescence in the range of 1100-1200 nm from SiNWs on their length was found for the first time. The reported study was funded by RFBR according to the research project No. 14-02-31544 мол_a.

Resume : During last decade, nanostructured semiconductors have raised as a promising route to fabrication of miniature chip-based Thermoelectric (TE) devices. Among them, Si has emerged as a potential TE candidate since the discovery that small diameter NWs conduct heat like a disordered solid, while maintaining reasonable values for both electrical conductivity and Seebeck coefficient if they remain crystalline, making the doping process of such structures a crucial step. Although the fabrication of miniature chips formed by Si NWs arrays may yield efficient conversion devices, the lack of reproducibility of bottom-up strategies difficult their establishment. Here, we present geometry based in Si thin films that also takes profit of the phonons scattering while easies the integration with planar technologies. As mentioned before a critical step in the fabrication process is to obtain the optimal doping level in the Si thin films. We opted to implant phosphorous ions (P+) and boron ions (B+) by plasma. To prevent effusion of dopant during post-implantation thermal treatments employed for re-crystallization, the Si thin films were capped by silicon nitride. To obtain the appropriate doping level (~10^19 at/cm^3 for TE) and with the aim to preserve the crystalline seed at the bottom of the thin films, detailed TRIM simulations, multiple implantation and annealing experiments were carried out. Withe final device, the TE properties of the p-n Si thin film couples are measured.

Resume : Multi-Gate Junctionless MOSFETs have emerged over the last years as promising candidates for low cost nanoelectronic applications, due to the superior electrical and fabrication process properties offered by these devices for both in digital as well as in analog applications. However, the use of uniformly doped channel, source and drain regions presents the well-known problem of the high series resistance associated to the extensions, which degrades the electrical performance of the device. Therefore, new designs and accurate models of nanoscale DGJ MOSFET including the defects at the interface Si/SiO2 are required for the comprehension of the fundamentals of such device behavior against the ageing phenomenon. Based on 2D numerical investigation of a nanoscale DGJ MOSFET, in the present work a numerical study for I-V and small signal characteristics by including both the highly doped extension regions and the interfacial defects is presented. The investigated design, which is a technologically feasible technique by introducing only one ion implantation step, provides a good solution to improve the device immunity against the interfacial defects. In this context, I-V and analog characteristics are investigated by an appropriate 2-D numerical modeling, where the obtained results are compared with those of the conventional DGJ MOSFET.