Search

Monolithic tandem cells involving a top cell with Si nanocrystals embedded in SiC (Si NC/SiC) and a c-Si bottom cell have been prepared. Scanning electron microscopy shows that the intended cell architecture is achieved and that it survives the 1100 degrees C anneal required to form Si NCs. The cells exhibit mean open-circuit voltages V-oc of 900-950mV, demonstrating tandem cell functionality, with 580mV arising from the c-Si bottom cell and 320mV arising from the Si NC/SiC top cell. The cells are successfully connected using a SiC/Si tunnelling recombination junction that results in very little voltage loss. The short-circuit current densities j(sc) are, at 0.8-0.9 mAcm(-2), rather low and found to be limited by current collection in the top cell. However, equivalent circuit simulations demonstrate that in current-mismatched tandem cells such as the ones studied here, higher j(sc), when accompanied by decreased V-oc, can arise from shunts or breakdown in the limiting cell rather than improved current collection from the limiting cell. This indicates that V-oc is a better optimisation parameter than j(sc) for tandem cells where the limiting cell exhibits poor junction characteristics. The high-temperature-stable cell architecture developed in this work, coupled with simulations highlighting potential pitfalls in tandem cell analysis, provides a suitable route for optimisation of Si NC layers for photovoltaics on a tandem cell device level. Copyright (c) 2016 John Wiley & Sons, Ltd.

The influence of temperature on the open-circuit voltage (VOC) of crystalline silicon solar cells is analysed using different semiconductor temperature models with different levels of accuracy. The strongest influence besides the direct dependence of the intrinsic carrier concentration on temperature results from the temperature dependence of the band gap and the effective density of states, while the incomplete ionization plays a minor role for the implied voltage. However incomplete ionization can play an important role for the external voltage at temperatures below 50K due to imperfect selectivity of the emitter and back surface field. The observed saturation of VOC towards low temperatures is caused by the effective density of states. The temperature dependence from 80 K to 300 K and the intensity dependence as a function of temperature and illumination density were measured on a silicon wafer solar cell resulting in a maximum voltage of 1012 mV at T=85.8K. The measured values could be well described by theory.

Si NCs embedded in a wide band gap material are promising candidates for all-Si tandem solar cells thanks to the possibility of tuning the NC band gap by varying their size, offered by the NC quantum confinement. The high temperature treatment required for Si crystallization represents a concern for the integration of this material in a tandem solar cell production route. The limiting step for solid phase crystallization is grain nucleation. A high temperature nucleation seeding followed by grain growth at a lower temperature is therefore a promising method to reduce the thermal budget. To this aim, we report on different annealing routes pursued to decrease the thermal budget: in all cases the germs of the c-Si nucleation are created by a high temperature (1150 °C) spike, then the crystallization follows at different temperatures, above and below the 3C-SiC solid phase epitaxial growth temperature (800 °C). The results are discussed in terms of the structural features produced by each process: the time required for complete dehydrogenation, mandatory to avoid disruptive hydrogen evolution during the high temperature spike; Si and SiC crystallization; defect nucleation in c-SiC, correlated to the rate of temperature rise during the annealing; role of the c-Si/SiC interface energy in affecting the growth process. The photovoltaic properties of the resulting structures are estimated based on the measured optical absorption, electronic transitions, and electrical transport.

Ordered silicon nanocrystals in silicon carbide are produced by Plasma Enhanced Chemical Vapor Deposition by means of the multilayer approach followed by annealing at 1100 °C. The crystallization is verified by Raman scattering, X-ray diffraction, Transmission Electron Microscopy, and UV–vis spectroscopy. The conditions for the periodic structure to survive the high temperature annealing and for the SiC barrier to confine the Si crystal growth are examined by energy-filtered transmission electron microscopy and X-ray reflection. The final layout appears to be strongly influenced by the structural features of the as-deposited multilayer. Threshold values of Si-rich carbide sublayer thickness and Si-to-C ratio are identified in order to preserve the ordered structure. The crystallized fraction is observed to be correlated with the total silicon volume fraction. The constraints are examined through the use of ab-initio calculations of matrix-embedded silicon nanocrystals, as well as in terms of existing models for nanocrystal formation, in order to establish the role played by the interface energy on nanocrystal outgrowth, residual amorphous fraction, and continuous crystallization. A parameter space of formation of ordered Si nanocrystals is proposed. The diffusivity of carbon in the crystallized material is evaluated, and estimated to be around 10–16 cm2/s at 1100 °C.

Silicon nanocrystal quantum dots in a dielectric matrix form a material with higher band gap than silicon, but still compatible with silicon technology. So far, devices using silicon nanocrystals have been realized either on silicon wafers, or using in-situ doping in the superlattice deposition which may hinder the nanocrystal formation. In this paper, a vertical PIN device is presented which allows to investigated the electrical and photovoltaic properties of nanocrystal quantum dot layers. The device structure circumvents any influence of a substrate wafer or dopants and provides full flexibility in the material choice of both, i.e. electron and hole, contacts. Furthermore, not-high-temperature stable contact materials can be applied. Devices have been realized using SiC/Si nanocrystal multilayers as the i-region and doped a-SixC1-x:H layers as electron and hole contacts. First devices show open-circuit voltage of up to 400mV.

Charge transport and electroluminescence mechanisms in Si-rich Si oxynitride/silicon oxide (SRON/SiO2) superlattices deposited on p-type Si substrate are reported. The superlattice structures were deposited by plasma-enhanced chemical-vapor deposition and subsequently annealed at 1150 °C to precipitate and crystallize the Si excess into Si nanocrystals. The dependence of the electrical conduction on the applied voltage and temperature was found to be well described by a Poole-Frenkel transport mechanism over a wide voltage range. On the other hand, the observed dependence of the electroluminescence on the SRON layer thickness is a clear proof of quantum confinement and was attributed to an excitonic radiative recombination taking place in the confined states within the Si quantum dots. A model is proposed based on thermal hopping of electrons between the quantum dots acting as trap states (Poole-Frenkel). A correlation between carrier transport and electroluminescence has been established considering impa...

Charge transport in nanocrystalline SiC with and without embedded Si nanocrystals (Si NCs) prepared by annealing of $a\ensuremath{-}\mathrm{S}{\mathrm{i}}_{1\ensuremath{-}x}{\mathrm{C}}_{x}:\mathrm{H}$ precursors is studied using temperature-dependent current-voltage measurements supported by electron spin resonance and mass spectrometry data. Transport is Ohmic in all films at all temperatures and the temperature dependence of conductivity shows that the materials behave as disordered semiconductors, exhibiting extended-state transport at high temperature and variable-range hopping transport at low temperature. Grain-boundary-, surface-, and interface-dominated transport is systematically ruled out. Films are $n$ type, and films with Si NCs exhibit up to ${10}^{3}$ times higher conductivity (up to $0.1\phantom{\rule{0.16em}{0ex}}\mathrm{S}\phantom{\rule{0.16em}{0ex}}\mathrm{c}{\mathrm{m}}^{\ensuremath{-}1}$) after exposure to a hydrogen plasma which passivates dangling bonds, particularly on Si NCs. A forming gas anneal does not have such an effect, indicating that atomic rather than molecular hydrogen is required. The conductivity of SiC films without Si NCs is largely unchanged by passivation and the Fermi level is not raised nearly as closely to the conduction band. This is attributed to a type I band offset between Si NCs and SiC that leads to extended-state conduction in films with Si NCs taking place in a Si network. This is confirmed by the dependence of the extended-state mobility on the volume fraction of excess Si. Variable-range hopping is relatively insensitive to the presence of excess Si and is hence considered to take place via shallow defect states throughout the volume of the films. The high conductivities are found to be a consequence of background doping by oxygen and nitrogen.

Nowadays, silicon solar cells have been the most widespread for commercial use. One of the most promising structure for further improvement of photoelectric conversion efficiency is tandem structure, which connects two types of solar cell. Recently, efficiency of perovskite solar cell (PSCs) has reached over 20%1) in few years, while silicon solar cell has reported as 26.3% for long time2). From fig.1, PSC has a large band gap of 1.5 eV, which makes open circuit voltage (V oc) high. Therefore, it is desirable to apply to a tandem top cell with silicon bottom cell. Fig.2 shows the concept of this tandem solar cell. Ideally, this structure can obtain over 30% efficiency3), but it is required to fabricate thin perovskite film with large grain as an absorber layer. Since a certain range of the wavelength exists in which both perovskite and silicon absorb, we need to balance a photoabsorption between the top and bottom cell by controlling the thickness of perovskite layer for the current matching between each layer. In conventional solar cell, an absorber layer which has a small grain size with many boundary causes fast recombination. Here, the larger grain size makes less grain boundary can reduce this effect. Generally, the growth of larger grain takes a long time, therefore it is needed to prevent following two negative effects: (1) discontinuity due to crystal growth toward 3-dimentional (3D) and (2) the decomposition of the perovskite layer. In most cases, solution process was adopted to fabricate the perovskite layer, which dissolved precursor of lead iodide (PbI2) and methylammonium iodide (MAI) in an organic solvent. However, the crystals from solution process grow in 3D and large grains were achieved, resulting in the low coverage (appearing the discontinuity). Also, film morphology of the perovskite is severely affected by drying process of the solution. Interestingly, another possible approach to make thin and continuous perovskite film is a dry process. After depositing PbI2, the deposited PbI2 reacted with MAI over 150°C by using sublimation. Most of the report, this process conducted at 150°C4). This process will keep PbI2morphology at some extent and 2D crystal growth can be expected. Higher temperature annealing is expected to be more effective to obtain a larger crystal grain, but the CH3NH3PbI3 perovskite is easy to decompose when annealing at high temperature and even MAI decomposition starts at 230-250°C5). Accordingly, there are only few reports conducted at a high temperature in this process. In this study, we investigated high-temperature annealing by two kinds of dry process (1.vapor deposition 2. Solid-state reaction) to obtain large grain toward 2D crystal growth. Furthermore, we discussed the effect of high-temperature annealing at different temperature and MAI concentration. In vapor deposition process, PbI2 layer was deposited on the planar TiO2 layer, and on the other side, MAI was deposited by spin coating MAI solution on another glass. Then, both substrates put face to face sandwiching with thermostable tape as a spacer. When the temperature increased from 150 to 180°C, the grain size of the perovskite was much larger than that of PbI2. However, the maximum grain size was limited to 1μm. Photoabsorbance of perovskite increased with increase annealing temperature, but some PbI2 was still detected in this temperature, even though the reaction time extended to about 2h. It indicated that the extent of reaction is still low at this temperature. When temperature reached to 200°C, reaction of PbI2 completed, but the perovskite layer detached from the substrate. It is supposed that excess MAI was segregated after completely reacted with PbI2. In the solid state reaction, MAI powder was put on the deposited PbI2 and directly annealed at 250°C. PbI2 was completely reacted and grain size was larger, but it was still limited to 1μm. Some pores were observed in the interface between TiO2 and the perovskite. It is supposed that maximum grain size of the perovskite might be limited to the grain structure of PbI2. Thus, it is important to obtain the suitable morphology of PbI2 for further improvement of the perovskite grain size. [References] 1) Yang. W et al., science, 2015, 348, 1234-1237 2) M. Green et al., Prog. Photovolt: Res. Appl. 2017, 2016, 25, 3-13 3) P. Löper et al., Phys. Chem. Chem. Phys., 2015, 17, 1619-1629 4) Liu. M et al., Nature, 2013, 501, 395-398 5) Emilio. J et al., Energy Environ. Sci., 2016, 9, 3406-3410 Figure 1

The diffusion of boron in nanocrystalline silicon carbide (nc-SiC) films with a grain size of 4-7 nm is studied using a poly-Si boron source. Diffusion is found to be much faster than in monocrystalline SiC as it takes place within the grain boundary (GB) network. Drive-in temperatures of 900-1000 °C are suitable for creating shallow boron profiles up to 100 nm deep, while 1100 °C is sufficient to flood the 200 nm thick films with boron. From the resulting plateau at 1100 °C a boron segregation coefficient of 28 between nc-SiC and the Si substrate, as well as a GB boron solubility limit of 0.2 nm−2 is determined. GB diffusion in the bulk of the films is Fickian and thermally activated with DGB(T)=(3.1−5.6)×107exp(−5.03±0.16 eV/kBT) cm2s−1. The activation energy is interpreted in terms of a trapping mechanism at dangling bonds. Higher boron concentrations are present at the nc-SiC surface and are attributed to immobilized boron.

28th European Photovoltaic Solar Energy Conference and Exhibition; 142-146, We present the first crystalline Si-based tandem solar cell device consisting of a monolithically interconnected c-Si wafer bottom and a Si nanocrystal (NC) top solar cell. The major part of this work has been realised within a European consortium in the frame of the FP7 project NASCEnT. Within this project we have been able to (i) develop new nanomaterials with photovoltaic compatible technologies, to (ii) improve the theoretical understanding of transport, recombination and dopant incorporation in these quantum confined absorber materials and to (iii) design devices which enable us to prove quantum confinement. In this publication we show that although solid phase crystallisation at 1100 °C for around 30 min is necessary to generate Si NCs of proper quality the wafer bottom cell as well as the c-SiC(n+) tunnel recombination junction layer can withstand such harsh conditions. Besides simulations to determine the influence of the high thermal budget on dopant profiles in wafer and c-SiC lifetime measurements on parallel processed wafers still show minority carrier lifetimes of 600 μs. The bifacial character of the tandem device enables us to clearly distinguish between bottom and top cell contributions. With a NC layer thickness of initially 3 nm Voc values as high as 978 mV could be achieved at 1 sun illumination from the front-side. Under rear-side illumination the short circuit current Jsc is with 1.0 mA/cm2 4 times higher that under front side illumination due to the better absorption and carrier lifetime of the bottom cell but still limited by the series resistance of the intrinsic NC absorber material. However, at 510 mV the Voc is much lower, indicating the difference in Voc is due to photogenerated voltage in the NC top cell which can only be created by illumination with low-wavelength light, i.e. front illumination. This very first tandem solar cell device including quantum dot absorber layers produced with photovoltaic compatible processes is the next step to successfully overcome the Shockley-Queisser limit for wafer based Si solar cells.

Silicon nanocrystals embedded in SiC are studied by spectrophotometry and photoluminescence (PL) spectroscopy. Absorptivities are found to be affected by residual Fabry-Pérot interference arising from measurements of reflection and transmission at locations of different film thickness. Multiple computational and experimental methods to avoid these errors in thin film measurements, in general, are discussed. Corrected absorptivity depends on the quantity of Si embedded in the SiC but is independent of the Si crystallinity, indicating a relaxation of the k-conservation criterion for optical transitions in the nanocrystals. Tauc gaps of 1.8-2.0 and 2.12 eV are determined for Si nanoclusters and SiC, respectively. PL spectra exhibit a red-shift of ∼100 nm per nm nominal Si nanocluster diameter, which is in agreement with quantum confinement but revealed to be an artifact entirely due to Fabry-Pérot interference. Several simple experimental methods to diagnose or avoid interference in PL measurements are developed that are applicable to all thin films. Corrected PL is rather weak and invariant with passivation, indicating that non-paramagnetic defects are responsible for rapid non-radiative recombination. They are also responsible for the broad, sub-gap PL of the SiC, and can wholly account for the form of the PL of samples with Si nanoclusters. The PL intensity of samples with Si nanoclusters, however, can only be explained with an increased density of luminescent defects in the SiC due to Si nanoclusters, efficient tunneling of photogenerated carriers from Si nanoclusters to SiC defects, or with emission from a-Si nanoclusters. Films prepared on Si exhibit much weaker PL than the same films prepared on quartz substrates.